Apparatus and method for monitoring and testing sensitive operations systems

Abstract

The invention relates to an apparatus for monitoring of an external system, wherein the external system comprises an operating component, wherein the apparatus comprises a monitoring component, wherein the monitoring component is configured to analyze the components of the operating component and / or at least one data wherein the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component, wherein the operating component is configured to execute at least one operation wherein the at least one operation is monitored by the monitoring component The invention also relates to a method for monitoring of an external system, wherein the external system is configured to operate an operating component, wherein the method comprises operating a monitoring component, wherein operating the monitoring component comprises analyzing the components of the operating component and / or at least one data wherein operating the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component, wherein operating the operating component comprises executing at least one operation wherein the at least one operation is monitored by the monitoring component.

Description

[0001] Apparatus and Method for monitoring and testing sensitive operations systems

[0002] Field

[0003] The invention lies in the field of monitoring and testing devices. More particularly, the invention lies in the field of monitoring and testing devices related to sensitive operations systems such as, but not limited to autonomous systems, and critical infrastructures.

[0004] Introduction

[0005] Monitoring systems play a vital role in ensuring the safety, reliability, and efficiency of autonomous systems and critical infrastructure by continuously observing their inputs and outputs. These systems analyze incoming data, such as sensor readings, environmental conditions, and user commands, while simultaneously tracking outputs, including system actions, performance metrics, and operational outcomes. By monitoring both ends, they can detect anomalies, predict failures, and ensure that operations align with expected behaviors. In autonomous systems, such as self-driving cars or drones, this dual observation enhances real-time decision-making and resilience. Similarly, in critical infrastructure like power grids, transportation networks, and water systems, it safeguards against disruptions, cyberattacks, and inefficiencies. With the integration of technologies like artificial intelligence and the Internet of Things (loT), modern monitoring systems are increasingly capable of identifying issues.

[0006] Development such as these have been put to use in the following inventions:

[0007] US11580604B1 depicts methods and systems for monitoring use and determining risks associated with operation of a vehicle having one or more autonomous operation features are provided. According to certain aspects, operating data may be recorded during operation of the vehicle. This may include information regarding the vehicle, the vehicle environment, use of the autonomous operation features, and / or control decisions made by the features. The control decisions may include actions the feature would have taken to control the vehicle, but which were not taken because a vehicle operator was controlling the relevant aspect of vehicle operation at the time. The operating data may be recorded in a log, which may then be used to determine risk levels associated with vehicle operation based upon risk levels associated with the autonomous operation features. The risk levels may further be used to adjust an insurance policy associated with the vehicle.

[0008] EP4365779A1 depicts a method, systems and devices for monitoring performance data of an algorithm and / or artificial intelligence, Al, model deployed on a platform; determining one or more performance criteria for the algorithm and / or Al model; comparing the monitored performance data to the one or more performance criteria; triggering an alert event based on the comparison of the monitored data to the one or more performance criteria; characterized in that the one or more performance criteria are determined based on a predetermined set of rules provided by a central server communicatively coupled to the platform, the set of rules being associated with the algorithm, the Al model and / or the platform.

[0009] However, despite their advanced capabilities, most current monitoring systems in autonomous systems and critical infrastructure focus solely on observing inputs and outputs, rather than examining the internal operations to isolate issues and respond accordingly. This means they primarily detect deviations in data or performance at the boundaries of the system but lack insight into the underlying processes driving those deviations. Without access to the internal mechanisms, such systems often struggle to identify root causes of anomalies or failures, relying instead on surface-level patterns to trigger alerts or actions. As a result, their ability to generate targeted corrective measures is limited, potentially leading to inefficiencies or incomplete resolutions in complex, multilayered systems.

[0010] Summary

[0011] In light of the above, it is therefore an object of the present invention to overcome or at least to alleviate the shortcomings and disadvantages of the prior art. More particularly, it is an object of the present invention to address the inefficiencies of the prior art, especially related to the internal monitoring of autonomous and critical infrastructures.

[0012] These objects are met by the present invention.

[0013] In a first aspect, the invention may relate to an apparatus for monitoring of an external system wherein the apparatus may comprise a monitoring component. The external system may comprise an operating component. The invention also may relate to an apparatus wherein the apparatus may also comprise an operating component, and a monitoring component. The operating component may be configured to execute at least one operation wherein the at least one operation may be monitored by the monitoring component. The operating component may further comprise a navigation component. The apparatus may additionally and / or alternatively comprise an acquiring component and / or a processing component. The modules of the invention may comprise one or more processing units configured to carry out computer instructions of a program (i.e. machine readable and executable instructions). The processing unit(s) may be singular or plural. For example, the data- processing system may comprise at least one of CPU, GPU, TPU, DSP, APU, ASIC, ASIP or FPGA. The data processing system may comprise memory components, such as, main memory (e.g. RAM), cache memory (e.g. SRAM) and / or secondary memory (e.g. HDD, SDD). The data processing system may comprise volatile and / or non-volatile memory such an SDRAM, DRAM, SRAM, Flash Memory, MRAM, F-RAM, or P-RAM.

[0014] Furthermore, the apparatus's components may be configured for analog computing and / or mixed computing. At least one of the apparatus's components may also be configured for analog computing and / or mixed computing. The apparatus's components may further be configured for neuromorphic computing. Additionally and / or alternatively, at least one of the apparatus's components may be configured for neuromorphic computing.

[0015] Moreover, the modules of the invention may comprise one or more analog and / or neuromorphic processing units configured to carry out computer instructions of a program (i.e. machine readable and executable instructions). The analog and / or neuromorphic processing unit(s) may be singular or plural. For example, the evaluation module may comprise at least one of but not limited to analog matrix processor(s), analog modular processor(s), reconfigurable analog modular processor(s), analog chips making use of but not limited to memristors, or neuromorphic chip(s).

[0016] Analog computing, is defined in this document as a type of computation where continuous physical phenomena, such as electrical voltage, mechanical motion, or fluid dynamics, are used to model and solve problems. Analog computing may refer to but is not limited to neuromorphic computing, defined as an approach to computing that mimics the structure and function of the human brain designing hardware and systems that replicate the neural networks found in biological brains. Analog and neuromorphic computing present better energy efficiency, faster processing and reduced complexity compared to regular computing. The implementation of analog and / or neuromorphic computing presents a preferred advantage of the current invention.

[0017] Additionally, neuromorphic computing and / or simulation of neuromorphic data originating data passing through the system according to embodiments cited herein through a processing unit may be used to prevent and / or alleviate a bottle neck of data created by the writing speed of the hardware configured for storing data going through the system according to any embodiments cited herein. Neuromorphic computing and / or the simulation of neuromorphic data would reduce the amount of data created to analyze, to be analyzed, to be sent, and / or to be stored. Such an implementation presents a further advantage.

[0018] Moreover, the apparatus may be configured for the deployment monitoring of a decisionmaking system. The apparatus may also be configured for the deployment monitoring of an external system. Deployment monitoring may be defined as continuously tracking and assessing the performance, behavior, and health of an external system after it has been deployed to production or other environments. The deployment monitoring may comprise but may not be limited to monitoring of transport components, decision-making components with an impact on safety and / or well-being, object detection. For example, dangerous objects' detection might fail by making flawed decisions in wrongly classifying an object or by wrongly predicting the objects movements, and hence will be monitored to avoid these flawed decisions. The deployment monitoring may also comprise but may not be limited to monitoring a wind turbine or a bridge which might have failed by falsely classifying an objects state, such as healthy (in good condition), and hence will be monitored to avoid these flawed decisions. The deployment monitoring may hence relate to monitoring of detection, classification, forecasting and other predictions tasks.

[0019] The apparatus may further be an apparatus for deployment monitoring of an unmanned autonomous machine, an aircraft and / or aircraft system, a watercraft and / or watercraft system, a vessel, a drone, and / or at least one critical infrastructure.

[0020] Critical infrastructure may be defined as an asset, a facility, equipment, a network or a system, or a part of an asset, a facility, equipment, a network or a system, which is necessary for the provision of an essential service, incorporated herein by reference to Art 2, point (4) of Directive (EU) 2022 / 2557 as of 14 December 2022. Critical infrastructure may include but may not be limited to roads, railways, bridges and transport systems, energy generation (gas / hydrogen / oil pipelines, wind turbines etc.), energy transmission, distribution and relay, food supply (on ships for example), telecommunications, such as underwater internet cables, medical supply (delivery drones for delivering medicine), smart devices for real-time consumption prediction and load balancing in a power grid or similar, at least partly autonomous drones operating close to power lines or energy facilities for inspection or similar purposes, smart devices for water, air and / or food quality monitoring where wrong alarms might cut water supply temporarily with fatal consequences, smart systems for traffic flow control and traffic management, seismic sensor systems for early- detection of earth quakes for protection of critical infrastructure (might apply to similar disaster management tasks).

[0021] A critical infrastructure may also comprise at least one high risk Al system, wherein a high- risk Al system may be defined as an Al system is intended to be used as a safety component of a product, or the Al system is itself a product, covered by the Union harmonisation legislation; and the product whose safety component is the Al system, or the Al system itself as a product, is required to undergo a third-party conformity assessment, with a view to the placing on the market or the putting into service of that product pursuant to the Union harmonisation legislation, wherein the Union harmonization legislation is comprised in Annex I of the EU artificial intelligence act. The definition of a high-risk Al system would be incorporated by reference to Part I, Art 6. a of the EU artificial intelligence act as of 12thof July 2024.

[0022] High-risk Al systems may comprise but may not be limited to Public services, Law Enforcement, any Al system that helps services such as but not limited to officers or public servants with situational awareness (marking citizens as suspects or as dangerous) which might lead to deprivation of liberty, detention, prolonged questioning, smart body cams for police officers or on-site drug testing or testing for explosives, pattern recognition ai systems which mark citizens as suspicious or similar, intelligent cyber-security, content monitoring, IT forensics tools that help to decide on suspicious content, activities, users which might lead to wrong accusation of citizens if not working properly, safety components of products, AI systems for injury prevention in industrial workshops, smart-home carbon monoxide or smoke detectors, airbag systems or smart braking algorithms in cars or other vehicles.

[0023] The apparatus may additionally and / or alternatively be configured for testing purposes. The apparatus may thus be configured to test an external system and / or the external system to check if it abides by desired standards. Standards may comprise but may not be limited to measurement standards related to data accessed and / or generated by the apparatus, certification standards related to certification, laws, and / or norms.

[0024] In one embodiment, the acquiring component may be configured to acquire at least one data and / or at least one metadata related to acquire the at least one data. For example, at least one metadata may comprise the resolution of the at least one data. The apparatus may be configured to analyze the impact of the at least one data according to the at least one metadata and / or detect at least one flaw of the operating component according to the at least one metadata. In this example, the method may comprise analyzing the impact of low quality data (either inaccurate data or imprecise data, i.e. high noise or wrong mean) and / or revealing the flaws of an ai caused by low quality data, i.e. cases in which it may be not immediately obvious the AI was wrong.

[0025] The acquiring component may also comprise at least one sensor, wherein the at least one sensor may comprise at least one sensor such as but not limited to at least one IMU, at least one barometric altitude sensor, at least one optical sensor, and / or at least one GNSS receiver. The at least on sensor may further be configured to acquire at least one data and / or at least one metadata related to the at least one data. The at least one data may be acquired in real time.

[0026] The at least one sensor may be configured as a digital sensor, an analog sensor, and / or a neuromorphic sensor. The acquiring component may also comprise a multitude of sensors, wherein the multitude of sensors may be configured to acquire at least one data. The multitude of sensors may be configured as any combination of digital sensors, analog sensors and / or neuromorphic sensors.

[0027] In the case of using neuromorphic data, it should be obvious to a person skilled in the art that the data will be represented in a way that neuromorphic devices may interpret. For example, an image data would not comprise colored pixels like pixels comprised in an image captured by a regular camera, but would comprise the difference in movement captured by the neuromorphic camera, in that example, and represented in a grid corresponding to the resolution of the neuromorphic camera.

[0028] Furthermore, the at least one data may relate to at least one data such as but not limited to at least one propulsion data wherein the at least one propulsion data may comprise at least one locomotion data, at least one motion data and / or at least one navigation data. The at least one data may also relate to at least one data such as but not limited to at least one LiDAR (Light Detection and Ranging) point data, at least one 3D point data, wherein the at least one 3D point data may comprise at least one spatial data, at least one position data, at least one velocity data, at least one time data, at least one atmospheric pressure data, at least one angular motion, at least one acceleration, at least one drive voltage, at least one rotational speed, at least one temperature data, and / or at least one weather data.

[0029] Moreover, the acquiring component may be configured to transmit the at least one data to the processing component. The acquiring component may also be configured to acquire at least one data from at least one component of the operating component. The acquiring component may also be configured to acquire at least one input, throughput and / or output of at least one component of the operating component.

[0030] In another embodiment, the processing component may be configured to process at least one data into at least one processed data. The processing component may also be configured to process the at least one data according to the at least one sensor. The processing component may further be configured to synchronize the at least one data with respect to other at least another data and / or with respect to time.

[0031] The processing component may be configured to filter out at least one outlier that may be comprised in the at least one data. The processing component may also be configured to identify at least one extremum in the at least one data. The processing component may further be configured to convert at least one measurement unit according to the at least one data. The processing component may additionally and / or alternatively be configured to transpose the at least one data from one coordinate system to another coordinate system.

[0032] Furthermore, the processing component may be configured to implement at least one frequency matching algorithm, wherein the at least one frequency matching algorithm may be configured to find at least one dominating frequency of at least one signal in the at least one data and / or process the at least one signal according to the at least one dominating frequency. The at least one frequency matching algorithm may comprise algorithms and / or techniques such as but not limited to STFT (short-time Fourier transform), and MFCC (Mel- frequency cepstral coefficients).

[0033] Moreover, the processing component may be configured for neuromorphic computing. The processing component may also be configured to convert the at least one data acquired by the at least one digital sensor and / or at least one analog sensor, such that the at least one data may be analyzed by a neuromorphic chip. The processing component may further be configured to convert the at least one data acquired by the at least one digital sensor and / or at least one analog sensor, such that the at least one data may be correctly analyzed by a neuromorphic chip.

[0034] In a further embodiment, the operating component may be configured to output at least one operational command wherein the at least one operational command may be configured to affect the apparatus and / or external system. The operating component may also comprise at least one Al-Operating component. The Al-operating component may be configured to output at least one operational command by making use of at least one Al algorithm, wherein the at least one operational command may be configured to affect the apparatus and / or external system. The at least one operational command may comprise but may not be limited to at least one warning, at least one navigation command, at least one display commands, and / or at least one command for the purpose of fulfilling at least one task.

[0035] Furthermore, the navigation component may be configured to output at least one navigational command, wherein the at least one navigational command may be configured to affect the position and / or orientation of the apparatus and / or the external system. The navigation component may also comprise a mission optimization component, wherein the mission optimization component may be configured to optimize at least one navigational command. The mission optimization component may further be configured to optimize at least one navigational command according to the at least one data.

[0036] Moreover, the operating component may comprise at least one extraction component, wherein the extraction component may be configured to extract at least one extracted data from at least one data. The extraction component may be configured to disaggregate at least one data.

[0037] In one embodiment, the monitoring component may be configured to analyze the components of the operating component and / or at least one data wherein the at least one data may be configured as input, throughput, and / or output according to each component may comprised in the operating component. The at least one data configured as throughput may refer to the at least one data going through the operating component. As such, throughput data may refer to data relating to internal workings of the operating component.

[0038] Furthermore, the monitoring component may be configured to classify at least one component of the operating component and / or at least one component of the apparatus into a trustworthy state and / or an untrustworthy state. A component in a trustworthy state may imply a reliable component, fulfilling task goals. A component in an untrustworthy state may imply an unreliable component, failing to meet task goals.

[0039] The monitoring component may be configured to compare at least one output of the operating component with a previous output of the operation component.

[0040] The wording used would be independent of testing or monitoring purposes.

[0041] The monitoring component may be configured to classify at least one data related to at least one component of the operating component, into at least one state according to the at least one data and at least one threshold, and / or according to at least one data related to the at least one component and a trusted reference.

[0042] The monitoring component may also be configured to classify the at least one component according to a representation of the at least one data and a representation of the trusted reference. A distance may refer to a measure, a norm, a difference, a divergence or a metric. For example, the apparatus may rely on vector distance, the distance may be between two latent vectors (such as the representation of signal now and a trusted representation). Another example would be the similarity measure "cosine similarity" related the angle between vectors, and quite advantageous in artificial intelligence / machine learning applications due its nice behavior in higher dimensions The monitoring component may further be configured to classify the at least one component according to the distance between the statistical distribution of the at least one data and the statistical distribution of the trusted reference, and / or according to the distance between the statistical distribution over time of the at least one data and the statistical distribution over time of the trusted reference. For example, the apparatus may be configured to compare the distribution of values (input / throughput / output) over some time interval and to what we would expect it to be, comparable to hypothesis / significance testing. The distribution of values may comprise multivariate distributions; during deployment, values (input / throughput / output) may co-occur (joint probability distribution) and the apparatus may thus be configured to determine the amount of change between co-occurences of values. The apparatus may also be configured to detect feature / covariate shift / drift in the Al-operating component.

[0043] The apparatus may then, for the example of monitoring an autonoumous flight component, for a lifetime metric that is updated after each flight, examine the erosion of the safety margin. Furthermore, the apparatus may calculate, for example, a projected arrival at safety margin based on a calculated erosion rate. The apparatus may then compute a metric based on an increasing variance of results over several uses of the operating component and / or flight.

[0044] The monitoring component may additionally and / or alternatively be configured to classify the at least one component according to the correlation between the at least one data and the trusted reference.

[0045] Moreover, the at least one data may relate to at least one metric, wherein at least two data points may relate to two different metrics. The monitoring component may also be configured to aggregate at least two metrics.

[0046] Additionally and / or alternatively, the monitoring component may be configured to improve the operating component. The monitoring component may also be configured to identify deficiencies relating to the operating component. The monitoring component may further be configured to modify at least one component of the operating component and / or at least one data related to at least one component of the operating component. In another embodiment, the monitoring component may comprise at least one AI- monitoring component. The Al-monitoring component may be configured to analyze the components of the operating component and / or at least one data wherein the at least one data may be configured as input, throughput, and / or output according to each component may comprised in the operating component by making use of at least one Al algorithm. The Al-monitoring component may also be configured to improve the Al-operating component. The Al-operating component and the Al-monitoring component may further be configured as a dual Al paradigm. A dual Al-algorithm may be defined as a framework or approach in artificial intelligence where two distinct Al systems or models may operate in a complementary manner, designed to leverage their respective strengths to achieve better performance, decision-making, or problem-solving capabilities. The apparatus may thus be configured to take advantage of the operating Al's ability to learn, and the monitoring Al's ability to affect the operating AL This interaction represents a preferred advantage of the described invention. The Al-monitoring component may comprise an ensemble of AIs, wherein the ensemble of AIs may be configured to perform the functions performed by the Al-monitoring component. For example, the ensemble of AIs may perform algorithms such as but not limited to bagging, boosting and / or voting.

[0047] The Al-operating component and the Al-monitoring component may be configured as at least one reinforcement learning algorithm. The Al-monitoring component and the AI- operating component may also be configured to implement an Actor-Critic algorithm, wherein the Al-operating component may be configured as the Actor and the Al-monitoring component may be configured as the Critic.

[0048] Furthermore, the Al-operating component and the Al-monitoring component may be configured as at least one generative algorithm. The Al-monitoring component and the AI- operating component may also be configured to implement a Generator-Discriminator algorithm, wherein the Al-operating component may be configured as the Generator and the Al-monitoring component may be configured as the Discriminator.

[0049] Moreover, the Al-operating component and the Al-monitoring component may be configured as at least one knowledge distillation algorithm. The Al-monitoring component and the Al-operating component may also be configured to implement a Student-Teacher algorithm, wherein the Al-operating component may be configured as the Student and the Al-monitoring component may be configured as the Teacher.

[0050] Additionally and / or alternatively, the Al-operating component and the Al-monitoring component may be configured as at least one transfer learning algorithm. The Al- monitoring component and the Al-operating component may also be configured to implement a Low-Rank Adaptation algorithm, wherein the Al-operating component may be configured as the specific-task model and the Al-monitoring component may be configured as the pre-trained model.

[0051] In one embodiment, the monitoring component may be configured to analyze the components of the operating component and / or at least one data wherein the at least one data may be configured as input according to each component may comprised in the operating component. The monitoring component may also be configured to analyze at least one component may comprised in the operating component and / or at least one data wherein the at least one data may be configured as input to the at least one component and / or to the operating component.

[0052] The monitoring component may be configured to execute at least one novelty rejection algorithm. The monitoring component may also be configured to execute at least one novelty rejection algorithm on the at least one data. The monitoring component may further be configured to calculate at least one anomaly score, wherein the at least one anomaly score may comprise a score based on at least one anomaly detected in at least one data. The monitoring component may additionally and / or alternatively be configured to calculate at least one anomaly score according to the result of the at least one novelty rejection algorithm. A novelty rejection algorithms may refer to but may not be limited to anomaly or outlier detection algorithms, break-point and change-point detection, and / or one-class classification.

[0053] In another embodiment, the monitoring component may be configured to analyze the components of the operating component and / or at least one data wherein the at least one data may be configured as throughput according to each component may comprised in the operating component. The monitoring component may also be configured to analyze at least one component may comprised in the operating component and / or at least one data wherein the at least one data may be configured as throughput to the at least one component and / or to the operating component.

[0054] The monitoring component may be configured to execute at least one novelty rejection algorithm on the at least one data. The monitoring component may also be configured to calculate at least one anomaly score according to the result of the at least one novelty rejection algorithm. The monitoring component may further be configured to execute at least one complexity reduction algorithm on at least one data, and / or at least one dimensionality reduction algorithm on at least one data. A complexity reduction algorithm may refer to but may not be limited to UMAP (Uniform Manifold Approximation and Projection) and / or its derivations. In a further embodiment, the monitoring component may be configured to analyze the components of the operating component and / or at least one data wherein the at least one data may be configured as output according to each component may comprised in the operating component. The monitoring component may also be configured to analyze at least one component may comprised in the operating component and / or at least one data wherein the at least one data may be configured as output to the at least one component and / or to the operating component. The monitoring component may further be configured to execute at least one novelty rejection algorithm on the at least one data. The monitoring component may additionally and / or alternatively be configured to calculate at least one anomaly score according to the result of the at least one novelty rejection algorithm.

[0055] Furthermore, the monitoring component may comprise a calibration component, wherein the calibration component may be configured to calibrate the operating component. The calibration component may also be configured to calibrate the operating component according to at least one throughput and / or output of the operating component, and / or according to the analysis of the monitoring component. The calibration component may further be configured to calibrate the operating component during deployment, and / or during testing. The calibration component may be configured to calibrate the operating component regardless of the type of task such as but not limited to classification tasks and / or regression tasks.

[0056] Moreover, the monitoring component may comprise at least one constraints component, wherein the constraints component may comprise at least one constraint that the operating component must abide by. The constraints component may also comprise at least one geographically dependent constraint. Examples of geographically dependent constraint may comprise but may not be limited to area of application of specific tasks, and / or country borders. The constraints component may further be configured to generate the at least one threshold according to at least one constraint. The at least one state may also be dependent on the at least one threshold and / or the at least one constraint. The constraints component may also be configured to generate at least one constraint, from at least one data related to at least one past operation carried out by the apparatus and / or past data acquired by the apparatus.

[0057] Additionally and / or alternatively, the monitoring component may comprise a sensor error detection component, wherein the sensor error detection component may be configured to detect at least one sensor error, according to the at least one data.

[0058] Furthermore, the monitoring component may comprise an attack detection and counter component, wherein the attack detection and counter component may be configured to detect at least one unauthorized manipulation of any component of the operating component. The attack detection and counter component may also be configured to counter at least one unauthorized manipulation of any component of the operating component. An unauthorized manipulation may comprise but may not be limited to GPS spoofing, DoS (Denial of Service) attacks, and / or adversarial Al attacks.

[0059] Moreover, the monitoring component may comprise a post-incident analysis component, wherein the post-incident analysis component may be configured to analyze at least one incident detected and / or countered by the attack detection and counter component. The post-incident analysis component may also be configured to transmit at least one analysis result to an external system and / or the external system. The post-incident analysis component may also be configured to transmit at least one data to an external incident analysis system.

[0060] Additionally and / or alternatively, the monitoring component may comprise a watchdog component, wherein the watchdog component may be configured to output at least one operational command, wherein the at least one operational command may comprise at least one warning signal. The watchdog component may also be configured to output at least one operational command wherein the at least one operational command may comprise at least one warning signal, according to the trustworthy state and / or untrustworthy state of the at least one component, and / or according to the at least one state.

[0061] In one embodiment, the apparatus may comprise a logger component, wherein the logger component may be configured as a tamper-proof storing device. The tamper proof property of the storing device may be implemented via hardware and / or software. The logger component may also be configured to store at least one data and / or at least one metadata The logger component may further be configured to store at least one data related to at least one past operation carried out by the apparatus and / or past data acquired by the apparatus. This past and / or historical data may be used to improve anomalies and identify anomalies that previously went undetected.

[0062] Logging may include processes such as but not limited to instrumentation involving implementing tools, code, or mechanisms to monitor, extract, and capture data from software or hardware components during execution, serialization involving converting data structures or objects into a standardized format for logging, storage, or transmission, storage and retention means for storing logs in appropriate storage systems with defined retention policies, anonymization processes in accordance with applicable legal or ethical standards, validation and integrity checks to ensure that log data is accurate, complete, and has not been tampered with.

[0063] The stored data may comprise structured, semi-structured, or unstructured data, from any of the internal components of the invention as described, interacting external systems, users, or external observers, whether human or automated systems operating outside the system boundaries. The stored data may further comprise timestamped events (such as but not limited to recorded events associated with a specific point in time, such as when a model generates a prediction, an error occurs, or a user interaction takes place,...), state snapshots (i.e., snapshots of the system state, such as memory usage, model status, or active components), inputs, throughputs, outputs, decisions (i.e., individual choices or actions taken by at least one Al implemented in the apparatus either autonomously or through human-in-the-loop mechanisms, e.g., approving a transaction or triggering an alarm), human oversight errors, error messages (i.e., warning messages or diagnostic records indicating instances of failure, exceptions, or problems in the operation of the system), environmental context (i.e., external conditions that may influence system behavior, such as network status, sensor readings, user loads, or events in the environment), and / or annotations (i.e., supplementary comments or metadata added manually or automatically that may describe system behavior, flag anomalies, or provide interpretive context). The stored data and metadata may be used by the system to identify any anomalies occurring within the operating component. The logger component may also be configured to record at least one data and / or metadata relating to at least one anomaly detected, such as but not limited to software errors, inputs / outputs with outlier values, potential attacks, external user system requests, inputs outside of the domain in which the operating component is supposed to work in, Al model shift and / or update, triggers from human oversight, missing data , incidents, accidents, near misses, short-term and longterm variations, and / or events (an event may be triggered externally resulting in the operating component making at least one decision and / or take at least one action). The stored data may comprise data configured to be reconstructed chronologically allowing for better testing and / or monitoring. The logger component may be configured to start logging data before and during the operation of the operating component. The stored data may comprise data modified specifically for the normal functioning of the operating component.

[0064] The at least one data and / or metadata may be used for training, calibrating and / or testing at least one component of the system. The stored data plays a role in the stability and robustness in the Al models implemented by the apparatus. The at least one stored data and / or metadata may enable support for continuous monitoring, trend analysis of the various variables, and / or accident and incident investigations. The logger component may be configured to selectively store the at least one data and / or metadata, allowing efficient usage of storage space available on the hardware comprised in the apparatus while still storing the at least one data and / or metadata required for the proper functioning of the monitoring component. The logger component may further be configured to selectively store at least one data according to at least one restraint and / or at least one threshold.

[0065] For example, events from streaming inputs or outputs may be logged at varying frequencies depending on the temporal resolution of the input, or at a frequency appropriate for monitoring a situation, such as at a higher frequency during a cyberattack.

[0066] Furthermore, any component of the apparatus may be configured to operate separately. The components of the apparatus may also be configured to communicate between each other, with the external system, and / or an external system. The apparatus may further be configured to train at least one Al algorithm, regardless of whether the at least one Al algorithm may be may comprised in the apparatus or not. The apparatus may also be configured to test at least one Al algorithm, regardless of whether the at least one Al algorithm may be may comprised in the apparatus or not.

[0067] Moreover, the monitoring component may comprise an operating divergence component, wherein the operating divergence component may be configured to detect at least one divergence in the functioning of the operating component and / or at least one component comprised in the operating component with respect to a normally functioning operating component and / or a normally functioning component comprised in the operating component. The operating divergence component may implement systems such as but not limited to Lyapunov systems. The operating divergence component may also be configured to detect at least one rate of divergence with respect to time of at least one divergence in the functioning of the operating component and / or at least one component comprised in the operating component, with respect to a normally functioning operating component and / or a normally functioning component comprised in the operating component. The operating divergence component may further be configured to detect at least one divergence in the functioning of the operating component according to at least one data related to at least one input and / or output to the operating component and / or the at least one component comprised in the operating component.

[0068] Additionally, the operating divergence component may be configured to detect at least one cyberattack and / or at least one implementation error. The operating divergence component may also be configured to detect at least one cyberattack and / or at least one implementation error occurring within the operation component according to the at least one rate of divergence. As the system may be autonomous, a component allowing detection of possible errors and / or attacks represents a further advantage in terms of identifying the trustworthiness of the operating component and eventually the amendment of the operating component as an autonomous system.

[0069] The Al-monitoring component may be configured to implement explainable Al (XAI) by making use of at least one data and / or metadata related to the at least one data, wherein the at least one data may be configured as input, throughput, and / or output according to each component that may be comprised in the operating component. This feature allows Al outputs to be traced, offering insights into the decision taken by an Al algorithm, allowing the Al-monitoring component to make use of such insight for the monitoring of the operating component, presenting a further advantage of the current invention.

[0070] The operating divergence component may be configured to output at least one signal to the operating component, wherein the at least one signal may be configured to relegate at least one automated operation and / or Al-based operation to a user-controlled system. The operating divergence component may also be configured to output at least one signal to the operating component, wherein the at least one signal may be configured to abort at least one operation operated by the operation component. These features allow for hybrid control of the operating component by a user-controlled system, presenting a further advantage of the current invention. An example of the interaction between a user- controlled system and the apparatus would relate to the apparatus recognizing human action as a second or third best option for operating the operating component and at the same time is certain that the best option is better, then the watchdog component may be configured to output a warning signal to the user-controlled system.

[0071] In a second aspect, the invention relates to a method for monitoring of an external system wherein the method may comprise operating a monitoring component. The external system may be configured to operate an operating component. The invention also relates to a method wherein the method may comprise operating an operating component, and operating a monitoring component. Operating the operating component may comprise executing at least one operation wherein the at least one operation may be monitored by the monitoring component. Operating the operating component may further comprise operating a navigation component. The method may additionally and / or alternatively comprise operating an acquiring component and / or a processing component. Operating the monitoring component would allow the autonomous operation of the operating component, such as but not limited to, in cases where data connections and / or communication with a user controlled system are not available. Furthermore, the components operated by the method may be configured for analog computing and / or mixed computing. At least one of the components operated by the method may also be configured for analog computing and / or mixed computing. The components operated by the method may further be configured for neuromorphic computing. Additionally and / or alternatively, at least one of the components operated by the method may be configured for neuromorphic computing.

[0072] Moreover, the method may be a method for the deployment monitoring of a decisionmaking system. The method may also be a method for the deployment monitoring of an external system. Deployment monitoring may be defined as continuously tracking and assessing the performance, behavior, and health of an external system after it has been deployed to production or other environments. The deployment monitoring may comprise but may not be limited to monitoring of transport components, decision-making components with an impact on safety and / or well-being, object detection. For example, dangerous objects' detection might fail by making flawed decisions in wrongly classifying an object or by wrongly predicting the objects movements, and hence will be monitored to avoid these flawed decisions. The deployment monitoring may also comprise but may not be limited to monitoring a wind turbine or a bridge which might have failed by falsely classifying an objects state, such as healthy (in good condition), and hence will be monitored to avoid these flawed decisions. The deployment monitoring may hence relate to monitoring of detection, classification, forecasting and other predictions tasks.

[0073] The method may further be a method for deployment monitoring of an unmanned autonomous machine, an aircraft and / or aircraft system, a watercraft and / or watercraft system, a vessel, a drone and / or at least one critical infrastructure.

[0074] Critical infrastructure may be defined as an asset, a facility, equipment, a network or a system, or a part of an asset, a facility, equipment, a network or a system, which is necessary for the provision of an essential service, incorporated herein by reference to Art 2, point (4) of Directive (EU) 2022 / 2557 as of 14 December 2022. Critical infrastructure may include but may not be limited to roads, railways, bridges and transport systems, energy generation (gas / hydrogen / oil pipelines, wind turbines etc.), energy transmission, distribution and relay, food supply (on ships for example), telecommunications, such as underwater internet cables, medical supply (delivery drones for delivering medicine), smart devices for real-time consumption prediction and load balancing in a power grid or similar, at least partly autonomous drones operating close to power lines or energy facilities for inspection or similar purposes, smart devices for water, air and / or food quality monitoring where wrong alarms might cut water supply temporarily with fatal consequences, smart systems for traffic flow control and traffic management, seismic sensor systems for early- detection of earth quakes for protection of critical infrastructure (might apply to similar disaster management tasks).

[0075] A critical infrastructure may also comprise high risk Al system, wherein a high-risk Al system may be defined as an Al system is intended to be used as a safety component of a product, or the Al system is itself a product, covered by the Union harmonisation legislation; and the product whose safety component is the Al system, or the Al system itself as a product, is required to undergo a third-party conformity assessment, with a view to the placing on the market or the putting into service of that product pursuant to the Union harmonisation legislation, wherein the Union harmonization legislation is comprised in Annex I of the EU artificial intelligence act. The definition of a high-risk Al system would be incorporated by reference to Part I, Art 6. a of the EU artificial intelligence act as of 12thof July 2024.

[0076] High-risk Al systems may comprise but may not be limited to Public services, Law Enforcement, any Al system that helps officers or public servants with situational awareness (marking citizens as suspects or as dangerous) which might lead to deprivation of liberty, detention, prolonged questioning, smart body cams for police officers or on-site drug testing or testing for explosives, pattern recognition ai systems which mark citizens as suspicious or similar, intelligent cyber-security, content monitoring, IT forensics tools that help to decide on suspicious content, activities, users which might lead to wrong accusation of citizens if not working properly, safety components of products, AI systems for injury prevention in industrial workshops, smart-home carbon monoxide or smoke detectors, airbag systems or smart braking algorithms in cars or other vehicles.

[0077] The method may additionally and / or alternatively be a method for testing purposes. The method may thus comprise testing an external system and / or the external system to check if it abides by desired standards. Standards may comprise but may not be limited to measurement standards related to data accessed and / or generated by the apparatus, certification standards related to certification, laws, and / or norms.

[0078] In one embodiment, operating the acquiring component may comprise acquiring at least one data and / or at least one metadata related to acquiring the at least one data. For example, at least one metadata may comprise the resolution of the at least one data. The method may comprise analyzing the impact of the at least one data according to the at least one metadata and / or detecting at least one flaw of the operating component according to the at least one metadata. In this example, the method may comprise analyzing the impact of low quality data (either inaccurate data or imprecise data, i.e. high noise or wrong mean) and / or revealing the flaws of an ai caused by low quality data, i.e. cases in which it may be not immediately obvious the AI was wrong.

[0079] Operating the acquiring component may also comprise operating at least one sensor, wherein operating the at least one sensor may comprise operating at least one sensor such as but not limited to at least one IMU, at least one barometric altitude sensor, at least one optical sensor, and / or at least one GNSS receiver. Operating the at least on sensor may further comprise acquiring at least one data and / or at least one metadata related to the at least one data. The at least one data may be acquired in real time.

[0080] The at least one sensor may be a digital sensor, an analog sensor, and / or a neuromorphic sensor. Operating the acquiring component may also comprise operating a multitude of sensors, wherein operating the multitude of sensors may comprise acquiring at least one data. Operating the multitude of sensors may comprise operating any combination of digital sensors, analog sensors and / or neuromorphic sensors.

[0081] In the case of using neuromorphic data, it should be obvious to a person skilled in the art that the data will be represented in a way that neuromorphic devices may interpret. For example, an image data would not comprise colored pixels like pixels comprised in an image captured by a regular camera, but would comprise the difference in movement captured by the neuromorphic camera, in that example, and represented in a grid corresponding to the resolution of the neuromorphic camera.

[0082] Furthermore, the at least one data may relate to at least one data such as but not limited to at least one propulsion data, wherein the at least one propulsion data may comprise at least one locomotion data, at least one motion data and / or at least one navigation data. The at least one data may also relate to at least one data such as but not limited to at least one LiDAR (Light Detection and Ranging) point data, at least one 3D point data, wherein the at least one 3D point data may comprise at least one spatial data, at least one position data, at least one velocity data, at least one time data, at least one atmospheric pressure data, at least one angular motion, at least one acceleration, at least one drive voltage, at least one rotational speed, at least one temperature data, and / or at least one weather data.

[0083] Moreover, operating the acquiring component may comprise transmitting the at least one data to the processing component. Operating the acquiring component may also comprise acquiring at least one data from at least one component of the operating component. Acquiring at least one data may comprise acquiring at least one input, throughput and / or output of at least one component of the operating component. In another embodiment, operating the processing component may comprise processing at least one data into at least one processed data. Operating the processing component may also comprise processing the at least one data according to the at least one sensor. Operating the processing component may further comprise synchronizing the at least one data with respect to other at least another data and / or with respect to time.

[0084] Operating the processing component may comprise filtering out at least one outlier that may comprised in the at least one data. Operating the processing component may also comprise identifying at least one extremum in the at least one data. Operating the processing component may further comprise converting at least one measurement unit according to the at least one data. Operating the processing component may additionally and / or alternatively comprise transposing the at least one data from one coordinate system to another coordinate system.

[0085] Furthermore, operating the processing component may comprise implementing at least one frequency matching algorithm, wherein implementing at least one frequency matching algorithm may comprise finding at least one dominating frequency of at least one signal in the at least one data and / or process the at least one signal according to the at least one dominating frequency. Implementing the at least one frequency matching algorithm may comprise implementing algorithms and / or techniques such as but not limited to STFT (short-time Fourier transform), and MFCC (Mel-frequency cepstral coefficients).

[0086] Moreover, the processing component operated by the method may be configured for neuromorphic computing. Operating the processing component may also comprise converting the at least one data acquired by the at least one digital sensor and / or at least one analog sensor, such that the at least one data may be analyzed by a neuromorphic chip. Operating the processing component may further comprise converting the at least one data acquired by the at least one digital sensor and / or at least one analog sensor, such that the at least one data may be correctly analyzed by a neuromorphic chip.

[0087] In a further embodiment, operating the operating component may comprise outputting at least one operational command wherein the at least one operational command may be configured to affect the external system and / or the components operated by the method. Operating the operating component may also comprise operating at least one Al-Operating component. Operating the Al-operating component may comprise outputting at least one operational command by making use of at least one Al algorithm, wherein the at least one operational command may be configured to affect the external system and / or the components operated by the method. The at least one operational command may comprise but may not be limited to at least one warning, at least one navigation command, at least one display commands, and / or at least one command for the purpose of fulfilling at least one task.

[0088] Furthermore, operating the navigation component may comprise outputting at least one navigational command, wherein the at least one navigational command may be configured to affect the position and / or orientation of the method and / or the external system. Operating the navigation component may also comprise operating a mission optimization component, wherein operating the mission optimization component may comprise optimizing at least one navigational command. Operating the mission optimization component may further comprise optimizing at least one navigational command according to the at least one data.

[0089] Moreover, operating the operating component may comprise operating at least one extraction component, wherein operating the extraction component may comprise extracting at least one extracted data from at least one data. Operating the extraction component may comprise disaggregating at least one data.

[0090] In one embodiment, operating the monitoring component may comprise analyzing the components of the operating component and / or at least one data wherein operating the at least one data may be configured as input, throughput, and / or output according to each component may comprised in the operating component. The at least one data configured as throughput may refer to the at least one data going through the operating component. As such, throughput data may refer to data relating to internal workings of the operating component.

[0091] Furthermore, operating the monitoring module may comprise classifying at least one component of the operating component and / or at least one component of the method into a trustworthy state and / or an untrustworthy state. A component in a trustworthy state may imply a reliable component, fulfilling task goals. A component in an untrustworthy state may imply an unreliable component, failing to meet task goals.

[0092] Operating the monitoring component may comprise comparing at least one output of the operating component with a previous output of the operation component.

[0093] Operating the monitoring component may comprise classifying at least one data related to at least one component of the operating component, into at least one state according to the at least one data and at least one threshold and / or according to at least one data related to the at least one component and a trusted reference. Operating the monitoring component may also comprise classifying the at least one component according to a representation of the at least one data and a representation of the trusted reference. A distance may refer to a measure, a norm, a difference, a divergence or a metric. For example, the method may rely on vector distance, the distance may be between two latent vectors (such as the representation of signal now and a trusted representation). Another example would be the similarity measure "cosine similarity" related the angle between vectors, and quite advantageous in artificial intelligence / machine learning applications due its nice behavior in higher dimensions. Operating the monitoring component may further comprise classifying the at least one component according to the distance between the statistical distribution of the at least one data and the statistical distribution of the trusted reference, and / or according to the distance between the statistical distribution over time of the at least one data and the statistical distribution over time of the trusted reference. For example, the method may comprise comparing the distribution of values (input / throughput / output) over some time interval and to what we would expect it to be, comparable to hypothesis / significance testing. The distribution of values may comprise multivariate distributions; during deployment, values (input / throughput / output) may cooccur (joint probability distribution) and the method may thus comprise determining the amount of change between co-occurences of values. The method may also comprise detecting feature / covariate shift / drift in the Al-operating component.

[0094] The method may then, for the example of monitoring an autonomous flight component, for a lifetime metric that is updated after each flight, comprise examining the erosion of the safety margin. Furthermore, the method may comprise calculating, for example, a projected arrival at safety margin based on a calculated erosion rate. The method may then comprise computing a metric based on an increasing variance of results over several uses of the operating component and / or flight.

[0095] Operating the monitoring component may additionally and / or alternatively comprise classifying the at least one component according to the correlation between the at least one data and the trusted reference.

[0096] Moreover, the at least one data may relate to at least one metric, wherein at least two data points may relate to two different metrics. Operating the monitoring component may also comprise aggregating at least two metrics.

[0097] Additionally and / or alternatively, operating the monitoring component may comprise improving the operating component. Operating the monitoring component may also comprise identifying deficiencies relating to the operating component. Operating the monitoring component may further comprise modifying at least one component of the operating component and / or at least one data related to at least one component of the operating component.

[0098] In another embodiment, operating the monitoring component may comprise operating at least one Al-monitoring component. Operating the Al-monitoring component may comprise analyzing the components of the operating component and / or at least one data wherein operating the at least one data may be configured as input, throughput, and / or output according to each component may comprised in the operating component by making use of at least one Al algorithm. Operating the Al-monitoring component may also comprise improving the Al-operating component. Operating the Al-operating component and operating the Al-monitoring component may further comprise operating a dual Al paradigm. A dual Al-algorithm may be defined as a framework or approach in artificial intelligence where two distinct Al systems or models may operate in a complementary manner, designed to leverage their respective strengths to achieve better performance, decision-making, or problem-solving capabilities. The method may thus comprise taking advantage of the operating Al's ability to learn and the monitoring Al's ability to affect the operating AL This interaction represents a preferred advantage of the described invention. Operating the Al-monitoring component may comprise operating an ensemble of AIs, wherein operating the ensemble of AIs comprise performing the functions performed by operating the Al-monitoring component. For example, the ensemble of AIs may perform algorithms such as but not limited to bagging, boosting and / or voting.

[0099] Operating the Al-operating component and operating the Al-monitoring component may comprise implementing at least one reinforcement learning algorithm. Operating the Al- monitoring component and operating the Al-operating component may also comprise implementing an Actor-Critic algorithm, wherein the Al-operating component may be configured as the Actor and the Al-monitoring component may be configured as the Critic.

[0100] Furthermore, operating the Al-operating component and operating the Al-monitoring component may comprise implementing at least one generative algorithm. Operating the Al-monitoring component and operating the Al-operating component may also comprise implementing a Generator-Discriminator algorithm, wherein the Al-operating component may be configured as the Generator and the Al-monitoring component may be configured as the Discriminator.

[0101] Moreover, operating the Al-operating component and operating the Al-monitoring component may comprise implementing at least one knowledge distillation algorithm. Operating the Al-operating component and operating the Al-monitoring component may also comprise implementing a Student-Teacher algorithm, wherein the Al-operating component may be configured as the Student and the Al-monitoring component may be configured as the Teacher.

[0102] Additionally and / or alternatively, operating the Al-operating component and operating the Al-monitoring component may comprise implementing at least one transfer learning algorithm. Operating the Al-operating component and operating the Al-monitoring component may comprise implementing a Low-Rank Adaptation algorithm, wherein the AI- operating component may be configured as the specific-task model and the Al-monitoring component may be configured as the pre-trained model.

[0103] In one embodiment, operating the monitoring component may comprise analyzing the components of the operating component and / or at least one data wherein the at least one data may be configured as input according to at least one component may comprised in the operating component operated by the method. Operating the monitoring component may also comprise analyzing at least one component may comprised in the operating component and / or at least one data wherein the at least one data may be configured as input to the at least one component and / or to the operating component operated by the method.

[0104] Operating the monitoring component may comprise executing at least one novelty rejection algorithm. Operating the monitoring component may also comprise executing at least one novelty rejection algorithm on the at least one data. Operating the monitoring component may further comprise calculating at least one anomaly score, wherein the at least one anomaly score may comprise operating a score based on at least one anomaly detected in at least one data. Operating the monitoring component may additionally and / or alternatively comprise calculating at least one anomaly score according to the result of the at least one novelty rejection algorithm. A novelty rejection algorithm may refer to but may not be limited to anomaly or outlier detection algorithms, break-point and changepoint detection, and / or one-class classification.

[0105] In another embodiment, operating the monitoring component may comprise analyzing the components of the operating component and / or at least one data wherein the at least one data may be configured as throughput according to at least one component may comprised in the operating component operated by the method. Operating the monitoring component may also comprise operating analyze at least one component may comprised in the operating component and / or at least one data wherein operating the at least one data may be configured as throughput to the at least one component and / or to the operating component operated by the method. Operating the monitoring component may comprise executing at least one novelty rejection algorithm on the at least one data. Operating the monitoring component may also comprise calculating at least one anomaly score according to the result of the at least one novelty rejection algorithm. Operating the monitoring component may further comprise executing at least one complexity reduction algorithm on at least one data, and / or at least one dimensionality reduction algorithm on at least one data. A complexity reduction algorithm may refer to but may not be limited to UMAP (Uniform Manifold Approximation and Projection) and / or its derivations.

[0106] In a further embodiment, operating the monitoring component may comprise analyzing the components of the operating component and / or at least one data wherein the at least one data may be configured as output according to at least one component may comprised in the operating component operated by the method. Operating the monitoring component may also comprise analyzing at least one component may comprised in the operating component and / or at least one data wherein the at least one data may be configured as output to the at least one component and / or to the operating component. Operating the monitoring component may further comprise executing at least one novelty rejection algorithm on the at least one data. Operating the monitoring component may additionally and / or alternatively comprise calculating at least one anomaly score according to the result of the at least one novelty rejection algorithm.

[0107] Furthermore, operating the monitoring component may comprise operating a calibration component, wherein operating the calibration component may comprise calibrating the operating component. Operating the calibration component may also comprise calibrating the operating component according to at least one throughput and / or output of the operating component, and / or according to the analysis of the monitoring component. Operating the calibration component may further comprise calibrating the operating component during deployment, and / or testing. Operating the calibration component may comprise calibrating the operating component regardless of the type of task such as but not limited to classification tasks and / or regression tasks.

[0108] Moreover, operating the monitoring component may comprise operating at least one constraints component, wherein operating the constraints component may comprise implementing at least one constraint that the operating component must abide by. Operating the constraints component may also comprise implementing at least one geographically dependent constraint. Examples of geographically dependent constraint may comprise but may not be limited to area of application of specific tasks, and / or country borders. Operating the constraints component may further comprise generating the at least one threshold according to at least one constraint. The at least one state may also be dependent on the at least one threshold and / or the at least one constraint. Operating the constraints component may also comprise generating at least one constraint and / or generating at least one constraint from at least one data related to at least one past operation carried out by the method and / or past data acquired by the method.

[0109] Additionally and / or alternatively, operating the monitoring component may comprise operating a sensor error detection component, wherein operating the sensor error detection component may comprise detecting at least one sensor error, according to the at least one data.

[0110] Furthermore, operating the monitoring component may comprise operating an attack detection and counter component, wherein operating the attack detection and counter component may comprise detecting at least one unauthorized manipulation of any component of the operating component. Operating the attack detection and counter component may further comprise operating counter at least one unauthorized manipulation of any component of the operating component. An unauthorized manipulation may comprise but may not be limited to GPS spoofing, DoS (Denial of Service) attacks, and / or adversarial Al attacks.

[0111] Moreover, operating the monitoring component may comprise operating a post-incident analysis component, wherein operating the post-incident analysis component may comprise analyzing at least one incident detected and / or countered by the attack detection and counter component. Operating the post-incident analysis component may also comprise transmitting at least one analysis result to an external system and / or the external system. Operating the post-incident analysis component may also comprise transmitting at least one data to an external incident analysis system.

[0112] Additionally and / or alternatively, operating the monitoring component may comprise operating a watchdog component, wherein operating the watchdog component may comprise outputting at least one operational command, wherein the at least one operational command may comprise at least one warning signal. Operating the watchdog component may also comprise outputting at least one operational command wherein the at least one operational command may comprise at least one warning signal, according to the trustworthy state and / or untrustworthy state of the at least one component, and / or according to the at least one state.

[0113] In one embodiment, the method may comprise operating a logger component, wherein operating the logger component may comprise operating a tamper-proof storing device. The tamper proof property of the storing device may be implemented via hardware and / or software. Operating the logger component may comprise storing at least one data, and / or at least one metadata. The at least one data and / or metadata may be used for training, calibrating and / or testing at least one component operated by the method. The stored data plays a role in the stability and robustness in the Al models implemented by the method. The stored data may also comprise data modified specifically for the normal functioning of the operating component and / or monitoring component. For example, the at least one data may be compressed for auditing purposes.

[0114] Operating the logger component may further comprise storing at least one data related to at least one past operation carried out by the method and / or past data acquired by the method. This past and / or historical data may be used to improve anomalies and identify anomalies that previously went undetected. Operating the logger component may also comprise selectively storing at least one data according to at least one restraint and / or at least one threshold.

[0115] Furthermore, operating any component operated by the method may comprise operating any component operated by the method independently from other components operated by the method. Operating any component operated by the method may also comprise communicating between each component operated by the method, the external system and / or an external system. The method may further comprise training at least one Al algorithm, regardless of whether the at least one Al algorithm may be may comprised in any component operated by the method or not. The method may also comprise testing at least one Al algorithm, regardless of whether the at least one Al algorithm may be may comprised in any component operated by the method or not.

[0116] Moreover, operating the monitoring component may comprise operating an operating divergence component, wherein operating the operating divergence component may comprise detecting at least one divergence in the operation of the operating component and / or at least one component comprised in the operating component with respect to the operation of a normally functioning operating component and / or a normally functioning component comprised in the operating component. Operating the operating divergence component may comprise implementing systems such as but not limited to Lyapunov systems. Operating the operating divergence component may also comprise detecting at least one rate of divergence with respect to time of at least one divergence in the operation of the operating component and / or at least one component comprised in the operating component, with respect to the operation of a normally functioning operating component and / or a normally functioning component comprised in the operating component. Operating the operating divergence component may further comprise detecting at least one divergence in the operation of the operating component according to at least one data related to at least one input and / or output to the operating component and / or the at least one component comprised in the operating component.

[0117] Additionally, operating the operating divergence component may comprise detecting at least one cyberattack and / or at least one implementation error. Operating the operating divergence component may also comprise detecting at least one cyberattack and / or at least one implementation error occurring with the operation of the operation component according to the at least one rate of divergence. As the method may involve operating an autonomous system, operating a component allowing detection of possible errors and / or attacks represents a further advantage in terms of identifying the trustworthiness of the operating component and eventually the amendment of the operating component as a method operating an autonomous system

[0118] Operating the Al-monitoring component comprises implementing explainable Al (XAI), by making use of at least one data and / or metadata related to the at least one data wherein the at least one data is configured as throughput according to at least one component comprised in the operating component operated by the method. This feature allows Al outputs to be traced, offering insights into the decision taken by an Al algorithm, allowing the operation of the Al-monitoring component to make use of such insight for the monitoring of the operating component, presenting a further advantage of the current invention.

[0119] Operating the operating divergence component may comprise outputting at least one signal to the operating component, wherein outputting the at least one signal may comprise relegating at least one automated operation and / or Al-based operation to a user- controlled system. Operating the operating divergence component may comprise outputting at least one signal to the operating component, wherein outputting the at least one signal may comprise aborting at least one operation operated by the operation component. These features allow for hybrid control of the operating component by a user- controlled system, presenting a further advantage of the current invention.

[0120] The method may be operated on the apparatus as described herein and vice versa. Below, apparatus embodiments will be discussed. These embodiments are abbreviated by the letter "S" followed by a number. When reference is herein made to a apparatus embodiment, those embodiments are meant.

[0121] 51. An apparatus for monitoring of an external system wherein the apparatus comprises a monitoring component.

[0122] 52. The apparatus according to the preceding apparatus embodiment wherein the external system comprises an operating component.

[0123] 53. An apparatus wherein the apparatus comprises an operating component, and a monitoring component.

[0124] 54. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiments S2 and / or S3, wherein the operating component is configured to execute at least one operation wherein the at least one operation is monitored by the monitoring component.

[0125] 55. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S2 and / or S3, wherein the operating component comprises a navigation component.

[0126] 56. The apparatus according to any of the preceding apparatus embodiment wherein the apparatus comprises an acquiring component.

[0127] 57. The apparatus according to any of the preceding apparatus embodiment wherein the apparatus comprises a processing component.

[0128] 58. The apparatus according to any preceding apparatus embodiment wherein the apparatus's components are configured for analog computing and / or mixed computing.

[0129] 59. The apparatus according to any preceding apparatus embodiment wherein at least one of the apparatus's components is configured for analog computing and / or mixed computing.

[0130] S10. The apparatus according to any preceding apparatus embodiment wherein the apparatus's components are configured for neuromorphic computing. 511. The apparatus according to any preceding apparatus embodiment wherein at least one of the apparatus's components is configured for neuromorphic computing.

[0131] 512. The apparatus according to any of the preceding apparatus embodiments wherein the apparatus is configured for the deployment monitoring of a decision-making system.

[0132] 513. The apparatus according to any preceding apparatus embodiment wherein the apparatus is configured for the deployment monitoring of an external system.

[0133] 514. The apparatus according to any preceding apparatus embodiment wherein the apparatus is an apparatus for deployment monitoring of an unmanned autonomous machine.

[0134] 515. The apparatus according to any preceding apparatus embodiment wherein the apparatus is an apparatus for deployment monitoring of an aircraft and / or aircraft system.

[0135] 516. The apparatus according to any preceding apparatus embodiment wherein the apparatus is an apparatus for deployment monitoring of a watercraft and / or watercraft system.

[0136] 517. The apparatus according to any preceding apparatus embodiment wherein the apparatus is an apparatus for deployment monitoring of a vessel.

[0137] 518. The apparatus according to any preceding apparatus embodiment wherein the apparatus is an apparatus for deployment monitoring of drones.

[0138] 519. The apparatus according to any preceding apparatus embodiment wherein the apparatus is an apparatus for monitoring at least one critical infrastructure.

[0139] 520. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiment S6, wherein the acquiring component is configured to acquire at least one data.

[0140] 521. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiment S20, wherein the acquiring component is configured to acquire at least one metadata related to acquire the at least one data.

[0141] 522. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S6, wherein the acquiring component comprises at least one sensor.

[0142] 523. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S22, wherein the at least one sensor comprises at least one IMU.

[0143] 524. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S22, wherein the at least one sensor comprises at least one barometric altitude sensor.

[0144] 525. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S22, wherein the at least one sensor comprises at least one optical sensor.

[0145] 526. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S22, wherein the at least one sensor comprises at least one GNSS receiver,

[0146] 527. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S22, wherein the at least on sensor is configured to acquire at least one data and / or at least one metadata related to the at least one data.

[0147] 528. The apparatus according to any preceding apparatus embodiments with the features of S20 and / or S27 wherein the at least one data is acquired in real time.

[0148] 529. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S22, wherein the at least one sensor is configured as a digital sensor.

[0149] 530. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S22, wherein the at least one sensor is configured as an analog sensor.

[0150] 531. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S22, wherein the at least one sensor is configured as a neuromorphic sensor.

[0151] 532. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S6, wherein the acquiring component is configured to comprise a multitude of sensors.

[0152] 533. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S32, wherein the multitude of sensors are configured to acquire at least one data.

[0153] 534. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S32, wherein the multitude of sensors is configured as any combination of digital sensors, analog sensors and / or neuromorphic sensors.

[0154] 535. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one propulsion data.

[0155] 536. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one LiDAR point data.

[0156] 537. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one 3D point data.

[0157] 538. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one position data.

[0158] 539. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one velocity data.

[0159] 540. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one time data. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one atmospheric pressure data. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one rotation rate. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one acceleration. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one drive voltage. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one rotational speed. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one temperature data. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S20 and / or S27, wherein the at least one data relates to at least one weather data. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S7 and S20, wherein the acquiring component is configured to transmit the at least one data to the processing component. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S2 or S3, and S6, wherein the acquiring component is configured to acquire at least one data from at least one component of the operating component. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S7 wherein the processing component is configured to process at least one data into at least one processed data. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S27 and S50, wherein the processing component is configured to process the at least one data according to the at least one sensor. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S50, wherein the processing component is configured to synchronize the at least one data with respect to other at least another data. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S50, wherein the processing component is configured to synchronize the at least one data with respect to time. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S50, wherein the processing component is configured to filter out at least one outlier comprised in the at least one data. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S50, wherein the processing component is configured to identify at least one extremum in the at least one data. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S50, wherein the processing component is configured to convert at least one measurement unit according to the at least one data. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S50, wherein the processing component is configured to transpose the at least one data from one coordinate system to another coordinate system. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S50, wherein the processing component is configured to implement at least one frequency matching algorithm, wherein the at least one frequency matching algorithm is configured to find at least one dominating frequency of at least one signal in the at least one data and / or process the at least one signal according to the at least one dominating frequency.

[0160] 559. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S7 and Sil, wherein the processing component is configured for neuromorphic computing.

[0161] 560. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S29 or S30, and S59, wherein the processing component is configured to convert the at least one data acquired by the at least one digital sensor and / or at least one analog sensor, such that the at least one data may be analyzed by a neuromorphic chip.

[0162] 561. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S29 or S30, and S59, wherein the processing component is configured to convert the at least one data acquired by the at least one digital sensor and / or at least one analog sensor, such that the at least one data may be correctly analyzed by a neuromorphic chip.

[0163] 562. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S2 and / or S3, wherein the operating component is configured to output at least one operational command wherein the at least one operational command is configured to affect the apparatus and / or external system.

[0164] 563. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S2 and / or S3, wherein the operating component comprises at least one Al-Operating component.

[0165] 564. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S63, wherein the Al-operating component is configured to output at least one operational command by making use of at least one Al algorithm.

[0166] 565. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S64, wherein the at least one operational command is configured to affect the apparatus and / or external system.

[0167] 566. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S5, wherein the navigation component is configured to output at least one navigational command.

[0168] 567. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S66, wherein the at least one navigational command is configured to affect the position and / or orientation of the apparatus and / or the external system.

[0169] 568. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S5, wherein the navigation component comprises a mission optimization component.

[0170] 569. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S68, wherein the mission optimization component is configured to optimize at least one navigational command.

[0171] 570. The apparatus according to any preceding apparatus embodiment with the features of any of apparatus embodiments S36-S41or S47, and the features of apparatus embodiment S68, wherein the mission optimization component is configured to optimize at least one navigational command according to the at least one data.

[0172] 571. The apparatus according to any preceding apparatus embodiment with the features of any of apparatus embodiments S2 or S3, wherein the operating component comprises at least one extraction component.

[0173] 572. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S71, wherein the extraction component is configured to extract at least one extracted data from at least one data.

[0174] 573. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S71, wherein the extraction component is configured to disaggregate at least one data.

[0175] 574. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to analyze the components of the operating component and / or at least one data wherein the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to classify at least one component of the operating component and / or at least one component of the apparatus into a trustworthy state and / or an untrustworthy state. The apparatus according to any preceding apparatus embodiment, wherein the monitoring component is configured to classify at least one data related to at least one component of the operating component, into at least one state according to the at least one data and at least one threshold. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S75, wherein the monitoring component is configured to classify the at least one component according to at least one data related to the at least one component and a trusted reference. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiment S77, wherein the monitoring component is configured to classify the at least one component according to a representation of the at least one data and a representation of the trusted reference. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiment S77, wherein the monitoring component is configured to classify the at least one component according to the distance between the statistical distribution of the at least one data and the statistical distribution of the trusted reference. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiment S77, wherein the monitoring component is configured to classify the at least one component according to the distance between the statistical distribution over time of the at least one data and the statistical distribution over time of the trusted reference. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiment S77, wherein the monitoring component is configured to classify the at least one component according to the correlation between the at least one data and the trusted reference. The apparatus according to any preceding apparatus embodiment wherein the at least one data relates to at least one metric.

[0176] 583. The apparatus according to any preceding apparatus embodiment wherein at least two data points relate to two different metrics.

[0177] 584. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S83, wherein the monitoring component is configured to aggregate at least two metrics.

[0178] 585. The apparatus according to any preceding apparatus embodiment, with the features of apparatus embodiments S2 and / or S3, wherein the monitoring component is configured to improve the operating component.

[0179] 586. The apparatus according to any preceding apparatus embodiment, with the features of apparatus embodiments S2 and / or S3, wherein the monitoring component is configured to identify deficiencies relating to the operating component.

[0180] 587. The apparatus according to any preceding apparatus embodiment, with the features of apparatus embodiments S2 and / or S3, wherein the monitoring component is configured to modify at least one component of the operating component and / or at least one data related to at least one component of the operating component.

[0181] 588. The apparatus according to any preceding apparatus embodiment wherein the monitoring component comprises at least one Al-monitoring component.

[0182] 589. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S88, wherein the Al-monitoring component is configured to analyze the components of the operating component and / or at least one data wherein the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component by making use of at least one Al algorithm.

[0183] 590. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S64 and S89 wherein the Al-monitoring component is configured to improve the Al-operating component.

[0184] 591. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S63 and S89 wherein the Al-operating component and the Al-monitoring component are configured as a dual Al paradigm.

[0185] 592. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S91, wherein the Al-operating component and the Al-monitoring component are configured as at least one reinforcement learning algorithm.

[0186] 593. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S92, wherein the Al-monitoring component and the Al-operating component are configured to implement an Actor-Critic algorithm, wherein the Al-operating component is configured as the Actor and the Al-monitoring component is configured as the Critic.

[0187] 594. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S91, the Al-operating component and the Al- monitoring component are configured as at least one generative algorithm.

[0188] 595. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S94, wherein the Al-monitoring component and the Al-operating component are configured to implement a Generator- Discriminator algorithm, wherein the Al-operating component is configured as the Generator and the Al-monitoring component is configured as the Discriminator.

[0189] 596. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S91, the Al-operating component and the Al- monitoring component are configured as at least one knowledge distillation algorithm.

[0190] 597. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S96, wherein the Al-monitoring component and the Al-operating component are configured to implement a Student-Teacher algorithm, wherein the Al-operating component is configured as the Student and the Al-monitoring component is configured as the Teacher.

[0191] 598. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S91, the Al-operating component and the Al- monitoring component are configured as at least one transfer learning algorithm. S99. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S98, wherein the Al-monitoring component and the Al-operating component are configured to implement a Low-Rank Adaptation algorithm, wherein the Al-operating component is configured as the specific-task model and the Al-monitoring component is configured as the pretrained model.

[0192] 5100. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to analyze the components of the operating component and / or at least one data wherein the at least one data is configured as input according to each component comprised in the operating component.

[0193] 5101. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to analyze at least one component comprised in the operating component and / or at least one data wherein the at least one data is configured as input to the at least one component and / or to the operating component.

[0194] 5102. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to execute at least one novelty rejection algorithm.

[0195] 5103. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S100 or S101 wherein the monitoring component is configured to execute at least one novelty rejection algorithm on the at least one data.

[0196] 5104. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to calculate at least one anomaly score, wherein the at least one anomaly score comprises a score based on at least one anomaly detected in at least one data.

[0197] 5105. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S103 and S104, wherein the monitoring component is configured to calculate at least one anomaly score according to the result of the at least one novelty rejection algorithm.

[0198] S106. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to analyze the components of the operating component and / or at least one data wherein the at least one data is configured as throughput according to each component comprised in the operating component.

[0199] 5107. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to analyze at least one component comprised in the operating component and / or at least one data wherein the at least one data is configured as throughput to the at least one component and / or to the operating component.

[0200] 5108. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S106 or S107 wherein the monitoring component is configured to execute at least one novelty rejection algorithm on the at least one data.

[0201] 5109. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S104 and S108, wherein the monitoring component is configured to calculate at least one anomaly score according to the result of the at least one novelty rejection algorithm.

[0202] Slid. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to execute at least one complexity reduction algorithm on at least one data.

[0203] 5111. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to execute at least one dimensionality reduction algorithm on at least one data.

[0204] 5112. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to analyze the components of the operating component and / or at least one data wherein the at least one data is configured as output according to each component comprised in the operating component.

[0205] 5113. The apparatus according to any preceding apparatus embodiment wherein the monitoring component is configured to analyze at least one component comprised in the operating component and / or at least one data wherein the at least one data is configured as output to the at least one component and / or to the operating component. 5114. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S112 or S113 wherein the monitoring component is configured to execute at least one novelty rejection algorithm on the at least one data.

[0206] 5115. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S104 and S114, wherein the monitoring component is configured to calculate at least one anomaly score according to the result of the at least one novelty rejection algorithm.

[0207] 5116. The apparatus according to any of the preceding apparatus embodiments wherein the monitoring component comprises a calibration component.

[0208] 5117. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiment S116, wherein the calibration component is configured to calibrate the operating component.

[0209] 5118. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiment S116, wherein the calibration component is configured to calibrate the operating component according to at least one throughput and / or output of the operating component.

[0210] 5119. The apparatus according to any of the preceding apparatus embodiments with the features of apparatus embodiments S74 and S116, wherein the calibration component is configured to calibrate the operating component according to the analysis of the monitoring component.

[0211] 5120. The apparatus according to any preceding apparatus embodiment with the features of any of apparatus embodiments S12-S18, and the features of apparatus embodiment S116, wherein the calibration component is configured to calibrate the operating component during deployment.

[0212] 5121. The apparatus according to any preceding apparatus embodiment with the features of any of apparatus embodiments with the features of apparatus embodiment S116, wherein the calibration component is configured to calibrate the operating component during testing.

[0213] S122. The apparatus according to any preceding apparatus embodiment wherein the monitoring component comprises at least one constraints component. S123. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S122, wherein the constraints component comprises at least one constraint that the operating component must abide by.

[0214] 5124. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S122, wherein the constraints component comprises at least one geographically dependent constraint.

[0215] 5125. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S76 and S122, wherein the constraints component is configured to generate the at least one threshold according to at least one constraint.

[0216] 5126. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S76 and S125, wherein the at least one state is dependent on the at least one threshold and / or the at least one constraint.

[0217] 5127. The apparatus according to any of the preceding embodiments, with the features of embodiment S122, wherein the constraints component is configured to generate at least one constraint.

[0218] 5128. The apparatus according to any of the preceding embodiments, with the features of embodiment S122, wherein the constraints component is configured to generate at least one constraint, from at least one data related to at least one past operation carried out by the apparatus and / or past data acquired by the apparatus.

[0219] 5129. The apparatus according to any preceding apparatus embodiment wherein the monitoring component comprises a sensor error detection component.

[0220] 5130. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S22 and S129, wherein the sensor error detection component is configured to detect at least one sensor error.

[0221] 5131. The apparatus according to any preceding apparatus embodiment with the features of any of apparatus embodiments S41, S42, S43, or S46, wherein the sensor error detection component is configured to detect at least one sensor error according to the at least one data. 5132. The apparatus according to any preceding apparatus embodiment wherein the monitoring component comprises an attack detection and counter component.

[0222] 5133. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S132, wherein the attack detection and counter component is configured to detect at least one unauthorized manipulation of any component of the operating component.

[0223] 5134. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S132, wherein the attack detection and counter component is configured to counter at least one unauthorized manipulation of any component of the operating component.

[0224] 5135. The apparatus according to any preceding apparatus embodiment wherein the monitoring component comprises a post-incident analysis component.

[0225] 5136. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S135, wherein the post-incident analysis component is configured to analyze at least one incident detected and / or countered by the attack detection and counter component.

[0226] 5137. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S136, wherein the post-incident analysis component is configured to transmit at least one analysis result to an external system and / or the external system.

[0227] 5138. The apparatus according to any preceding apparatus embodiment wherein the monitoring component comprises a watchdog component.

[0228] 5139. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S138, wherein the watchdog component is configured to output at least one operational command, wherein the at least one operational command comprises at least one warning signal.

[0229] 5140. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S75 and S138, wherein the watchdog component is configured to output at least one operational command wherein the at least one operational command comprises at least one warning signal, according to the trustworthy state and / or untrustworthy state of the at least one component.

[0230] 5141. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S76 and S138, wherein the watchdog component is configured to output at least one operational command wherein the at least one operational command comprises at least one warning signal, according to the at least one state.

[0231] 5142. The apparatus according to any preceding apparatus embodiment wherein the apparatus comprises a logger component.

[0232] 5143. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S142, wherein the logger component is configured as a tamper-proof storing device.

[0233] 5144. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S142, wherein the logger component is configured to store at least one data.

[0234] 5145. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S142, wherein the logger component is configured to store at least one metadata.

[0235] 5146. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiment S142, wherein the logger component is configured to store at least one data related to at least one past operation carried out by the apparatus and / or past data acquired by the apparatus.

[0236] 5147. The apparatus according to any preceding apparatus embodiment with the features of apparatus embodiments S142 and any of S76, S123-S128, wherein the logger component is configured to selectively store at least one data according to at least one restraint and / or at least one threshold.

[0237] 5148. The apparatus according to any preceding apparatus embodiment, wherein any component of the apparatus is configured to operate separately.

[0238] 5149. The apparatus according to any preceding apparatus embodiment, wherein the components of the apparatus are configured to communicate between each other. S150. The apparatus according to any preceding apparatus embodiment wherein the components of the apparatus are configured to communicate with the external system and / or an external system.

[0239] 5151. The apparatus according to any preceding apparatus embodiment, wherein the apparatus is configured to train at least one Al algorithm, regardless of whether the at least one Al algorithm is comprised in the apparatus or not.

[0240] 5152. The apparatus according to any preceding apparatus embodiments, wherein the apparatus is configured to test at least one Al algorithm, regardless of whether the at least one Al algorithm is comprised in the apparatus or not.

[0241] 5153. The apparatus according to any preceding apparatus embodiment, wherein the apparatus is configured for testing purposes.

[0242] 5154. The apparatus according to any of the preceding embodiments, wherein the monitoring component comprises an operating divergence component.

[0243] 5155. The apparatus according to any of the preceding embodiments, with the features of embodiments S4 and S154, wherein the operating divergence component is configured to detect at least one divergence in the functioning of the operating component and / or at least one component comprised in the operating component with respect to a normally functioning operating component and / or a normally functioning component comprised in the operating component.

[0244] 5156. The system according to any of the preceding embodiments, with the features of embodiments S4 and S154, wherein the operating divergence component is configured to detect at least one rate of divergence with respect to time of at least one divergence in the functioning of the operating component and / or at least one component comprised in the operating component, with respect to a normally functioning operating component and / or a normally functioning component comprised in the operating component.

[0245] 5157. The system according to any of the preceding embodiments with the features of embodiments S4 and S154 wherein the operating divergence component is configured to detect at least one divergence in the functioning of the operating component according to at least one data related to at least one input and / or output to the operating component and / or the at least one component comprised in the operating component. S158. The system according to any of the preceding embodiments, with the features of embodiment S154, wherein the operating divergence component is configured to detect at least one cyberattack and / or at least one implementation error.

[0246] 5159. The system according to any of the preceding embodiments, with the features of embodiment S156 and S158, wherein the operating divergence component is configured to detect at least one cyberattack and / or at least one implementation error occurring within the operation component according to the at least one rate of divergence.

[0247] 5160. The system according to any of the preceding embodiments, with the features of embodiment S88, wherein the Al-monitoring component is configured to implement explainable Al (XAI).

[0248] 5161. The system according to any of the preceding embodiments, with the features of embodiment S160, wherein the Al-monitoring component is configured to implement explainable Al (XAI) by making use of at least one data and / or metadata related to the at least one data, wherein the at least one data may be configured as input, throughput, and / or output according to each component may comprised in the operating component.

[0249] 5162. The system according to any of the preceding embodiments, with the features of embodiments S2 and / or S3, and S154, wherein the operating divergence component is configured to output at least one signal to the operating component, wherein the at least one signal is configured to relegate at least one automated operation and / or Al-based operation to a user-controlled system.

[0250] 5163. The system according to any of the preceding embodiments, with the features of embodiments S2 and / or S3, and S154, wherein the operating divergence component is configured to output at least one signal to the operating component, wherein the at least one signal is configured to abort at least one operation operated by the operation component.

[0251] 5164. The system according to any of the preceding embodiments, with the features of embodiments S2 and / or S3, wherein the monitoring component is configured to compare at least one output of the operating component with a previous output of the operation component. Below, method embodiments will be discussed. These embodiments are abbreviated by the letter "M" followed by a number. When reference is herein made to a method embodiment, those embodiments are meant.

[0252] Ml. A method for monitoring of an external system wherein the method comprises operating a monitoring component.

[0253] M2. The method according to the preceding method embodiment wherein the external system is configured to operate an operating component.

[0254] M3. A method wherein the method comprises operating an operating component, and operating a monitoring component.

[0255] M4. The method according to any of the preceding method embodiments with the features of method embodiments M2 and / or M3, wherein operating the operating component comprises executing at least one operation wherein the at least one operation is monitored by the monitoring component.

[0256] M5. The method according to any preceding method embodiment with the features of method embodiments M2 and / or M3, wherein operating the operating component comprises operating a navigation component.

[0257] M6. The method according to any of the preceding method embodiment wherein the method comprises operating an acquiring component.

[0258] M7. The method according to any of the preceding method embodiment wherein the method comprises operating a processing component.

[0259] M8. The method according to any preceding method embodiment wherein the components operated by the method are configured for analog computing and / or mixed computing.

[0260] M9. The method according to any preceding method embodiment wherein at least one of the components operated by the method is configured for analog computing and / or mixed computing. MIO. The method according to any preceding method embodiment wherein the components operated by the method are configured for neuromorphic computing.

[0261] Mil. The method according to any preceding method embodiment wherein at least one of the components operated by the method is configured for neuromorphic computing.

[0262] M12. The method according to any of the preceding method embodiments wherein the method is a method for the deployment monitoring of a decision-making system.

[0263] M13. The method according to any preceding method embodiment wherein the method is a method for the deployment monitoring of an external system.

[0264] M14. The method according to any preceding method embodiment wherein the method is a method for deployment monitoring of an unmanned autonomous machine.

[0265] M15. The method according to any preceding method embodiment wherein the method is a method for deployment monitoring of an aircraft and / or aircraft system.

[0266] M16. The method according to any preceding method embodiment wherein the method is a method for deployment monitoring of a watercraft and / or watercraft system.

[0267] M17. The method according to any preceding method embodiment wherein the method is a method for deployment monitoring of a vessel.

[0268] M18. The method according to any preceding method embodiment wherein the method is a method for deployment monitoring of drones.

[0269] M19. The method according to any preceding method embodiment wherein the method is a method for monitoring at least one critical infrastructure.

[0270] M20. The method according to any of the preceding method embodiments with the features of method embodiment M6, wherein operating the acquiring component comprises acquiring at least one data.

[0271] M21. The method according to any of the preceding method embodiments with the features of method embodiment M20, wherein operating the acquiring component comprises acquiring at least one metadata related to acquiring the at least one data.

[0272] M22. The method according to any preceding method embodiment with the features of method embodiment M6, wherein operating the acquiring component comprises operating at least one sensor.

[0273] M23. The method according to any preceding method embodiment with the features of method embodiment M22, wherein operating the at least one sensor comprises operating at least one IMU.

[0274] M24. The method according to any preceding method embodiment with the features of method embodiment M22, wherein operating the at least one sensor comprises operating at least one barometric altitude sensor.

[0275] M25. The method according to any preceding method embodiment with the features of method embodiment M22, wherein operating the at least one sensor comprises operating at least one optical sensor.

[0276] M26. The method according to any preceding method embodiment with the features of method embodiment M22, wherein operating the at least one sensor comprises operating at least one GNSS receiver,

[0277] M27. The method according to any preceding method embodiment with the features of method embodiment M22, wherein operating the at least on sensor comprises acquiring at least one data and / or at least one metadata related to the at least one data.

[0278] M28. The method according to any preceding method embodiments with the features of M20 and / or M27 wherein the at least one data is acquired in real time.

[0279] M29. The method according to any preceding method embodiment with the features of method embodiment M22, wherein the at least one sensor is a digital sensor.

[0280] M30. The method according to any preceding method embodiment with the features of method embodiment M22, wherein operating the at least one sensor is an analog sensor.

[0281] M31. The method according to any preceding method embodiment with the features of method embodiment M22, wherein operating the at least one sensor is a neuromorphic sensor.

[0282] M32. The method according to any preceding method embodiment with the features of method embodiment M6, wherein operating the acquiring component comprises operating a multitude of sensors.

[0283] M33. The method according to any preceding method embodiment with the features of method embodiment M32, wherein operating the multitude of sensors comprises acquiring at least one data.

[0284] M34. The method according to any preceding method embodiment with the features of method embodiment M32, wherein operating the multitude of sensors comprises operating any combination of digital sensors, analog sensors and / or neuromorphic sensors.

[0285] M35. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one propulsion data.

[0286] M36. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one LiDAR point data.

[0287] M37. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one 3D point data.

[0288] M38. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one position data.

[0289] M39. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one velocity data.

[0290] M40. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one time data. M41. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one atmospheric pressure data.

[0291] M42. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one rotation rate.

[0292] M43. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one acceleration.

[0293] M44. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one drive voltage.

[0294] M45. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one rotational speed.

[0295] M46. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one temperature data.

[0296] M47. The method according to any preceding method embodiment with the features of method embodiments M20 and / or M27, wherein the at least one data relates to at least one weather data.

[0297] M48. The method according to any preceding method embodiment with the features of method embodiments M7 and M20, wherein operating the acquiring component comprises transmitting the at least one data to the processing component.

[0298] M49. The method according to any preceding method embodiment with the features of method embodiments M2 or M3, and M6, wherein operating the acquiring component comprises acquiring at least one data from at least one component of the operating component.

[0299] M50. The method according to any preceding method embodiment with the features of method embodiment M7 wherein operating the processing component comprises processing at least one data into at least one processed data.

[0300] M51. The method according to any preceding method embodiment with the features of method embodiments M27 and M50, wherein operating the processing component comprises processing the at least one data according to the at least one sensor.

[0301] M52. The method according to any preceding method embodiment with the features of method embodiment M50, wherein operating the processing component comprises synchronizing the at least one data with respect to other at least another data.

[0302] M53. The method according to any preceding method embodiment with the features of method embodiment M50, wherein operating the processing component comprises synchronizing the at least one data with respect to time.

[0303] M54. The method according to any preceding method embodiment with the features of method embodiment M50, wherein operating the processing component comprises filtering out out at least one outlier comprised in the at least one data.

[0304] M55. The method according to any preceding method embodiment with the features of method embodiment M50, wherein operating the processing component comprises identifying at least one extremum in the at least one data.

[0305] M56. The method according to any preceding method embodiment with the features of method embodiment M50, wherein operating the processing component comprises converting at least one measurement unit according to the at least one data.

[0306] M57. The method according to any preceding method embodiment with the features of method embodiment M50, wherein operating the processing component comprises transposing the at least one data from one coordinate system to another coordinate system.

[0307] M58. The method according to any preceding method embodiment with the features of method embodiment M50, wherein operating the processing component comprises implementing at least one frequency matching algorithm, wherein operating the at least one frequency matching algorithm comprises finding at least one dominating frequency of at least one signal in the at least one data and / or process the at least one signal according to the at least one dominating frequency.

[0308] M59. The method according to any preceding method embodiment with the features of method embodiments M7 and Mil, wherein the processing component operated by the method is configured for neuromorphic computing.

[0309] M60. The method according to any preceding method embodiment with the features of method embodiments M29 or M30, and M59, wherein operating the processing component comprises converting the at least one data acquired by the at least one digital sensor and / or at least one analog sensor, such that the at least one data may be analyzed by a neuromorphic chip.

[0310] M61. The method according to any preceding method embodiment with the features of method embodiments M29 or M30, and M59, wherein operating the processing component comprises converting the at least one data acquired by the at least one digital sensor and / or at least one analog sensor, such that the at least one data may be correctly analyzed by a neuromorphic chip.

[0311] M62. The method according to any preceding method embodiment with the features of method embodiment M2 and / or M3, wherein operating the operating component comprises outputting at least one operational command wherein the at least one operational command is configured to affect the external system and / or the components operated by the method.

[0312] M63. The method according to any preceding method embodiment with the features of method embodiment M2 and / or M3, wherein operating the operating component comprises operating at least one Al-Operating component.

[0313] M64. The method according to any preceding method embodiment with the features of method embodiment M63, wherein operating the Al-operating component comprises outputting at least one operational command by making use of at least one Al algorithm.

[0314] M65. The method according to any preceding method embodiment with the features of method embodiment M64, wherein the at least one operational command is configured to affect the external system and / or the components operated by the method.

[0315] M66. The method according to any preceding method embodiment with the features of method embodiment M5, wherein operating the navigation component comprises outputting at least one navigational command. M67. The method according to any preceding method embodiment with the features of method embodiment M66, wherein the at least one navigational command is configured to affect the position and / or orientation of the method and / or the external system.

[0316] M68. The method according to any preceding method embodiment with the features of method embodiment M5, wherein operating the navigation component comprises operating a mission optimization component.

[0317] M69. The method according to any preceding method embodiment with the features of method embodiment M68, wherein operating the mission optimization component comprises optimizing at least one navigational command.

[0318] M70. The method according to any preceding method embodiment with the features of any of method embodiments M36-M41 or M47, and the features of method embodiment M68, wherein operating the mission optimization component comprises optimizing at least one navigational command according to the at least one data.

[0319] M71. The method according to any preceding method embodiment with the features of any of method embodiments M2 or M3, wherein operating the operating component comprises operating at least one extraction component.

[0320] M72. The method according to any preceding method embodiment with the features of method embodiment M71, wherein operating the extraction component comprises extracting at least one extracted data from at least one data.

[0321] M73. The method according to any preceding method embodiment with the features of method embodiment M71, wherein operating the extraction component comprises disaggregating at least one data.

[0322] M74. The method according to any preceding method embodiment wherein operating the monitoring component comprises analyzing the components of the operating component and / or at least one data wherein operating the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component.

[0323] M75. The method according to any preceding method embodiment wherein operating the monitoring component comprises classifying at least one component of the operating component and / or at least one component of the method into a trustworthy state and / or an untrustworthy state.

[0324] M76. The method according to any preceding method embodiment, wherein operating the monitoring component comprises classifying at least one data related to at least one component of the operating component, into at least one state according to the at least one data and at least one threshold.

[0325] M77. The method according to any preceding method embodiment with the features of method embodiment M75, wherein operating the monitoring component comprises classifying the at least one component according to at least one data related to the at least one component and a trusted reference.

[0326] M78. The method according to any of the preceding method embodiments with the features of method embodiment M77, wherein operating the monitoring component comprises classifying the at least one component according to a representation of the at least one data and a representation of the trusted reference.

[0327] M79. The method according to any of the preceding method embodiments with the features of method embodiment M77, wherein operating the monitoring component comprises classifying the at least one component according to the distance between the statistical distribution of the at least one data and the statistical distribution of the trusted reference.

[0328] M80. The method according to any of the preceding method embodiments with the features of method embodiment M77, wherein operating the monitoring component comprises classifying the at least one component according to the distance between the statistical distribution over time of the at least one data and the statistical distribution over time of the trusted reference.

[0329] M81. The method according to any of the preceding method embodiments with the features of method embodiment M77, wherein operating the monitoring component comprises classifying the at least one component according to the correlation between the at least one data and the trusted reference.

[0330] M82. The method according to any preceding method embodiment wherein the at least one data relates to at least one metric. M83. The method according to any preceding method embodiment wherein at least two data points relate to two different metrics.

[0331] M84. The method according to any preceding method embodiment with the features of method embodiment M83, wherein operating the monitoring component comprises aggregating at least two metrics.

[0332] M85. The method according to any preceding method embodiment, with the features of method embodiments M2 and / or M3, wherein operating the monitoring component comprises improving the operating component.

[0333] M86. The method according to any preceding method embodiment, with the features of method embodiments M2 and / or M3, wherein operating the monitoring component comprises identifying deficiencies relating to the operating component.

[0334] M87. The method according to any preceding method embodiment, with the features of method embodiments M2 and / or M3, wherein operating the monitoring component comprises modifying at least one component of the operating component and / or at least one data related to at least one component of the operating component.

[0335] M88. The method according to any preceding method embodiment wherein operating the monitoring component comprises operating at least one Al-monitoring component.

[0336] M89. The method according to any preceding method embodiment with the features of method embodiment M88, wherein operating the Al-monitoring component comprises analyzing the components of the operating component and / or at least one data wherein operating the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component by making use of at least one Al algorithm.

[0337] M90. The method according to any preceding method embodiment with the features of method embodiment M64 and M89 wherein operating the Al-monitoring component comprises improving the Al-operating component.

[0338] M91. The method according to any preceding method embodiment with the features of method embodiment M63 and M89 wherein operating the Al-operating component and operating the Al-monitoring component comprise operating a dual Al paradigm. M92. The method according to any preceding method embodiment with the features of method embodiment M91, wherein operating the Al-operating component and operating the Al-monitoring component comprise implementing at least one reinforcement learning algorithm.

[0339] M93. The method according to any preceding method embodiment with the features of method embodiment M92, wherein operating the Al-monitoring component and operating the Al-operating component comprise implementing an Actor-Critic algorithm, wherein the Al-operating component is configured as the Actor and the Al-monitoring component is configured as the Critic.

[0340] M94. The method according to any preceding method embodiment with the features of method embodiment M91 wherein operating the Al-operating component and operating the Al-monitoring component comprise implementing at least one generative algorithm.

[0341] M95. The method according to any preceding method embodiment with the features of method embodiment M94, wherein operating the Al-monitoring component and operating the Al-operating component comprise implementing a Generator- Discriminator algorithm, wherein the Al-operating component is configured as the Generator and the Al-monitoring component is configured as the Discriminator.

[0342] M96. The method according to any preceding method embodiment with the features of method embodiment M91, wherein operating the Al-operating component and operating the Al-monitoring component comprise implementing at least one knowledge distillation algorithm.

[0343] M97. The method according to any preceding method embodiment with the features of method embodiment M96, wherein operating the Al-operating component and operating the Al-monitoring component comprise implementing a Student- Teacher algorithm, wherein the Al-operating component is configured as the Student and the Al-monitoring component is configured as the Teacher.

[0344] M98. The method according to any preceding method embodiment with the features of method embodiment M91, wherein operating the Al-operating component and operating the Al-monitoring component comprise implementing at least one transfer learning algorithm. M99. The method according to any preceding method embodiment with the features of method embodiment M98, wherein operating the Al-operating component and operating the Al-monitoring component comprise implementing a Low-Rank Adaptation algorithm, wherein the Al-operating component is configured as the specific-task model and the Al-monitoring component is configured as the pretrained model.

[0345] M100. The method according to any preceding method embodiment wherein operating the monitoring component comprises analyzing the components of the operating component and / or at least one data wherein the at least one data is configured as input according to at least one component comprised in the operating component operated by the method.

[0346] Midi. The method according to any preceding method embodiment wherein operating the monitoring component comprises analyzing at least one component comprised in the operating component and / or at least one data wherein the at least one data is configured as input to the at least one component and / or to the operating component operated by the method.

[0347] M102. The method according to any preceding method embodiment wherein operating the monitoring component comprises executing at least one novelty rejection algorithm.

[0348] M103. The method according to any preceding method embodiment with the features of method embodiments M100 or Midi wherein operating the monitoring component comprises executing at least one novelty rejection algorithm on the at least one data.

[0349] Mld4. The method according to any preceding method embodiment wherein operating the monitoring component comprises calculating at least one anomaly score, wherein the at least one anomaly score comprises operating a score based on at least one anomaly detected in at least one data.

[0350] Mld5. The method according to any preceding method embodiment with the features of method embodiments Mld3 and Mld4, wherein operating the monitoring component comprises calculating at least one anomaly score according to the result of the at least one novelty rejection algorithm.

[0351] Mld6. The method according to any preceding method embodiment wherein operating the monitoring component comprises analyzing the components of the operating component and / or at least one data wherein the at least one data is configured as throughput according to at least one component comprised in the operating component operated by the method.

[0352] M107. The method according to any preceding method embodiment wherein operating the monitoring component comprises operating analyze at least one component comprised in the operating component and / or at least one data wherein operating the at least one data is configured as throughput to the at least one component and / or to the operating component operated by the method.

[0353] M108. The method according to any preceding method embodiment with the features of method embodiments M106 or M107 wherein operating the monitoring component comprises executing at least one novelty rejection algorithm on the at least one data.

[0354] M109. The method according to any preceding method embodiment with the features of method embodiments M104 and M108, wherein operating the monitoring component comprises calculating at least one anomaly score according to the result of the at least one novelty rejection algorithm.

[0355] MHO. The method according to any preceding method embodiment wherein operating the monitoring component comprises executing at least one complexity reduction algorithm on at least one data.

[0356] Mill. The method according to any preceding method embodiment wherein operating the monitoring component comprises executing at least one dimensionality reduction algorithm on at least one data.

[0357] M112. The method according to any preceding method embodiment wherein operating the monitoring component comprises analyzing the components of the operating component and / or at least one data wherein the at least one data is configured as output according to at least one component comprised in the operating component operated by the method.

[0358] M113. The method according to any preceding method embodiment wherein operating the monitoring component comprises analyzing at least one component comprised in the operating component and / or at least one data wherein the at least one data is configured as output to the at least one component and / or to the operating component.

[0359] M114. The method according to any preceding apparatus embodiment with the features of method embodiments M112 or M113 wherein operating the monitoring component comprises executing at least one novelty rejection algorithm on the at least one data.

[0360] M115. The method according to any preceding apparatus embodiment with the features of method embodiments M104 and M114, wherein operating the monitoring component comprises calculating at least one anomaly score according to the result of the at least one novelty rejection algorithm.

[0361] M116. The method according to any of the preceding method embodiments wherein operating the monitoring component comprises operating a calibration component.

[0362] M117. The method according to any of the preceding method embodiments with the features of method embodiment M116, wherein operating the calibration component comprises calibrating the operating component.

[0363] M118. The method according to any of the preceding method embodiments with the features of method embodiment M116, wherein operating the calibration component comprises calibrating the operating component according to at least one throughput and / or output of the operating component.

[0364] M119. The method according to any of the preceding method embodiments with the features of method embodiments M74 and M116, wherein operating the calibration component comprises calibrating the operating component according to the analysis of the monitoring component.

[0365] M120. The method according to any preceding method embodiment with the features of any of method embodiments M12-M18, and the features of method embodiment M116, wherein operating the calibration component comprises calibrating the operating component during deployment.

[0366] M121. The method according to any preceding method embodiment with the features of any of method embodiments with the features of method embodiment M116, wherein operating the calibration component comprises calibrating the operating component during testing. M122. The method according to any preceding method embodiment wherein operating the monitoring component comprises operating at least one constraints component.

[0367] M123. The method according to any preceding method embodiment with the features of method embodiment M122, wherein operating the constraints component comprises implementing at least one constraint that the operating component must abide by.

[0368] M124. The method according to any preceding method embodiment with the features of method embodiment M122, wherein operating the constraints component comprises implementing at least one geographically dependent constraint.

[0369] M125. The method according to any preceding method embodiment with the features of method embodiments M76 and M122, wherein operating the constraints component comprises generating the at least one threshold according to at least one constraint.

[0370] M126. The method according to any preceding method embodiment with the features of method embodiments M76 and M125, wherein the at least one state is dependent on the at least one threshold and / or the at least one constraint.

[0371] M127. The method according to any of the preceding method embodiments, with the features of embodiment M122, wherein operating the constraints component comprises generating at least one constraint.

[0372] M128. The method according to any of the preceding method embodiments, with the features of embodiment M122, wherein operating the constraints component comprises generating at least one constraint from at least one data related to at least one past operation carried out by the method and / or past data acquired by the method.

[0373] M129. The method according to any preceding method embodiment wherein operating the monitoring component comprises operating a sensor error detection component.

[0374] M130. The method according to any preceding method embodiment with the features of method embodiments M22 and M129, wherein operating the sensor error detection component comprises detecting at least one sensor error.

[0375] M131. The method according to any preceding method embodiment with the features of any of method embodiments M41, M42, M43, or M46, wherein operating the sensor error detection component comprises detecting at least one sensor error according to the at least one data.

[0376] M132. The method according to any preceding method embodiment wherein operating the monitoring component comprises operating an attack detection and counter component.

[0377] M133. The method according to any preceding method embodiment with the features of method embodiment M132, wherein operating the attack detection and counter component comprises detecting at least one unauthorized manipulation of any component of the operating component.

[0378] M134. The method according to any preceding method embodiment with the features of method embodiment M132, wherein operating the attack detection and counter component comprises operating counter at least one unauthorized manipulation of any component of the operating component.

[0379] M135. The method according to any preceding method embodiment wherein operating the monitoring component comprises operating a post-incident analysis component.

[0380] M136. The method according to any preceding method embodiment with the features of method embodiment M135, wherein operating the post-incident analysis component comprises analyzing at least one incident detected and / or countered by the attack detection and counter component.

[0381] M137. The method according to any preceding method embodiment with the features of method embodiment M136, wherein operating the post-incident analysis component comprises transmitting at least one analysis result to an external system and / or the external system.

[0382] M138. The method according to any preceding method embodiment wherein operating the monitoring component comprises operating a watchdog component.

[0383] M139. The method according to any preceding method embodiment with the features of method embodiment M138, wherein operating the watchdog component comprises outputting at least one operational command, wherein the at least one operational command comprises at least one warning signal.

[0384] M140. The method according to any preceding method embodiment with the features of method embodiment M75 and M138, wherein operating the watchdog component comprises outputting at least one operational command wherein the at least one operational command comprises at least one warning signal, according to the trustworthy state and / or untrustworthy state of the at least one component.

[0385] M141. The method according to any preceding method embodiment with the features of method embodiment M76 and M138, wherein operating the watchdog component comprises outputting at least one operational command wherein the at least one operational command comprises at least one warning signal, according to the at least one state.

[0386] M142. The method according to any preceding method embodiment wherein the method comprises operating a logger component.

[0387] M143. The method according to any preceding method embodiment with the features of method embodiment M142, wherein operating the logger component comprises operating a tamper-proof storing device.

[0388] M144. The method according to any preceding method embodiment with the features of method embodiment M142, wherein operating the logger component comprises storing at least one data.

[0389] M145. The method according to any preceding method embodiment with the features of method embodiment M142, wherein operating the logger component comprises storing at least one metadata.

[0390] M146. The method according to any preceding method embodiment with the features of method embodiment M142, wherein operating the logger component comprises storing at least one data related to at least one past operation carried out by the method and / or past data acquired by the method.

[0391] M147. The method according to any preceding method embodiment with the features of method embodiment M142 and any of M76, M123-M128, wherein operating the logger component comprises selectively storing at least one data according to at least one restraint and / or at least one threshold.

[0392] M148. The method according to any preceding method embodiment, wherein operating any component operated by the method comprises operating any component operated by the method independently from other components operated by the method.

[0393] M149. The method according to any preceding method embodiment, wherein operating any component operated by the method comprises communicating between each component operated by the method.

[0394] M150. The method according to any preceding method embodiment wherein operating any component operated by the method comprises communicating between each component operated by the method and the external system and / or an external system.

[0395] M151. The method according to any preceding method embodiment, wherein the method comprises training at least one Al algorithm, regardless of whether the at least one Al algorithm is comprised in any component operated by the method or not.

[0396] M152. The method according to any preceding method embodiments, wherein the method comprises testing at least one Al algorithm, regardless of whether the at least one Al algorithm is comprised in any component operated by the method or not.

[0397] M153. The method according to any preceding method embodiment, wherein the method is a method for testing purposes.

[0398] M154. The method according to any of the preceding method embodiments, wherein operating the monitoring component comprises operating an operating divergence component.

[0399] M155. The method according to any of the preceding method embodiments, with the features of embodiments M4 and M154, wherein operating the operating divergence component comprises detecting at least one divergence in the operation of the operating component and / or at least one component comprised in the operating component with respect to the operation of a normally functioning operating component and / or a normally functioning component comprised in the operating component. M156. The method according to any of the preceding method embodiments, with the features of embodiments M4 and M154, wherein operating the operating divergence component comprises detecting at least one rate of divergence with respect to time of at least one divergence in the operation of the operating component and / or at least one component comprised in the operating component, with respect to the operation of a normally functioning operating component and / or a normally functioning component comprised in the operating component.

[0400] M157. The method according to any of the preceding method embodiments with the features of embodiments M4 and M154 wherein operating the operating divergence component comprises detecting at least one divergence in the operation of the operating component according to at least one data related to at least one input and / or output to the operating component and / or the at least one component comprised in the operating component.

[0401] M158. The method according to any of the preceding method embodiments, with the features of embodiment S154, wherein operating the operating divergence component comprises detecting at least one cyberattack and / or at least one implementation error.

[0402] M159. The method according to any of the preceding method embodiments, with the features of embodiment S156 and S158, wherein operating the operating divergence component comprises detecting at least one cyberattack and / or at least one implementation error occurring with the operation of the operation component according to the at least one rate of divergence.

[0403] M160. The method according to any of the preceding method embodiments, with the features of embodiment M88, wherein operating the Al-monitoring component comprises implementing explainable Al (XAI).

[0404] M161. The method according to any of the preceding embodiments, with the features of embodiment M160, wherein operating the Al-monitoring component comprises implementing explainable Al (XAI) by making use of at least one data and / or metadata related to the at least one data, wherein the at least one data is configured as throughput according to at least one component comprised in the operating component operated by the method.

[0405] M162. The method according to any of the preceding embodiments, with the features of embodiments M2 and / or M3, and M154, wherein operating the operating divergence component comprises outputting at least one signal to the operating component, wherein outputting the at least one signal comprises relegating at least one automated operation and / or Al-based operation to a user-controlled system.

[0406] M163. The method according to any of the preceding embodiments, with the features of embodiments M2 and / or M3, and M154, wherein operating the operating divergence component comprises outputting at least one signal to the operating component, wherein outputting the at least one signal comprises aborting at least one operation operated by the operation component.

[0407] M164. The method according to any of the preceding embodiments, with the features of embodiments M2 and / or M3, wherein operating the monitoring component comprises comparing at least one output of the operating component with a previous output of the operation component.

[0408] Brief description of the figures

[0409] The present invention will now be described with reference to the accompanying drawings which illustrate embodiments of the invention. These embodiments should only exemplify, but not limit, the present invention.

[0410] Fig. 1 schematically depicts an example of the apparatus according to embodiments of the present invention;

[0411] Fig. 2 schematically depicts another example of the apparatus according to embodiments of the present invention;

[0412] Fig. 3 schematically depicts a further example of the apparatus according to embodiments of the present invention;

[0413] Fig. 4 schematically depicts an example of part of the apparatus according to embodiments of the present invention;

[0414] Fig. 5 depicts an example of the use of the apparatus according to embodiments of the present invention.

[0415] It is noted that not all the drawings carry all the reference signs. Instead, in some of the drawings, some of the reference signs have been omitted for sake of brevity and simplicity of illustration. Embodiments of the present invention will now be described with reference to the accompanying drawings.

[0416] Detailed description of the figures

[0417] Fig. 1 depicts the apparatus 1000, comprising the monitoring component 1100, interacting with an external system 2000, comprising an operating component 2200 performing an operation 2220. The input 2210 to the operation 2220 generates an output 2230. At least one of the internal values / data generated in operation 2220 to output the output 2230, are represented by the throughput 2225.

[0418] Monitoring component 1100 monitors the input 2210, the throughput 2225 and the output 2230 related to operation 2220. The monitoring component 1100 may output an operational command to the operating component 2200 such that the operational command relates to enhancing the performance, accuracy, and / or any performance measure of the operation 2220 and / or operation component 2200.

[0419] It should be obvious that the monitoring component may monitor multiple operations at once, and / or in real-time. Fig. 2 depicts the apparatus 1000, comprising the monitoring component 1100, an operating component 1200, an acquiring component 1300, a processing component 1400, and a logger component 1500. The operating component 1200 performing an operation 1220. The input 1210 to the operation 1220 generates an output 1230. At least one of the internal values / data generated in operation 1220 to output the output 1230, are represented by the throughput 1225.

[0420] The acquiring component 1300 may acquire the input 1210, the throughput 1225 and / or the output 1230 from the operation 1200. The acquiring may happen by making use of at least one sensor. The acquiring component may also acquire at least one data for the operation component 1200 such that the operating component 1200 makes use of the at least one data, regardless of the operating component's presence within the apparatus 1000 or an external system 2000.

[0421] The processing component 1400 may process the acquired data, the processing component may be configured to synchronize the data and / or clean the data obtained from the acquiring component 1300. The processing component may output the processed data to the operating component 1200 and / or the monitoring component 1100.

[0422] The logger component 1500 may store data and / or metadata related to the operating component, acquiring component and / or monitoring component. These data may also be transmitted to another external system 3000.

[0423] It should be obvious that the functions of the monitoring component 1100, the acquiring component 1300, the processing component 1400 and / or the logger component 1500 are not impacted by the location of the operating component 1200 with respect to the other components of the apparatus 1000, and / or are not impacted by the operating component's presence within the apparatus 1000 or an external system 2000.

[0424] The processing component 1400 may be comprised, at least in part, in the operating component 1200, the acquiring component 1300 and / or the monitoring component 1100. The acquiring component 1300 may be comprised at least in part in, the operating component 1200, the processing component 1400 and / or the monitoring component 1100.

[0425] Fig. 3 depicts the monitoring component 1100, the acquiring component 1300 and the operating component 1200 and / or 2200. The monitoring component 1100 comprises a calibration component 1120, a constraints component 1130, a watchdog component 1140, a sensor error detection component 1150, an attack detection and counter component 1160 and a post-incident analysis component. The monitoring component 1100 may monitor the operating component 1200 and / or 2200 by making use of at least one algorithm of the set of monitoring algorithms 1110. The at least one algorithm of the set of monitoring algorithms 1110 may make use of at least one Al-monitoring component 1115.

[0426] The set of monitoring algorithms 1110 and the operating component 1200 and / or 2200 operate as described previously in the description. The calibration component 1120 may make use of the output of at least one algorithm of the set of monitoring algorithms 1110 to calibrate the operation component 1200 and / or 2200.

[0427] The constraints component 1130 comprise at least one constraint that the operating component 1200 and / or 2200 must abide by. The monitoring component 1100 would then apply the at least one constraint to the operating component 1200 and / or 2200. Should any infraction of the at least one constraint occur, the constraints component 1130 may transmit an operational command to the watchdog component 1140 which will be configured to output at least one operational command wherein the at least one operation command may comprise at least one warning signal.

[0428] The watchdog component 1140 may also output an operating command wherein the operating command may comprise at least one warning signal wherein the at least one warning signal may relate to at least one state, and / or at least one trustworthy and / or untrustworthy state outputted by the set of monitoring algorithms 1110.

[0429] The sensor error detection component 1150 may be configured to detect at least one sensor error, wherein the sensor may be comprised in the acquiring component 1300. The sensor error component 1150 may be configured to send at least one operational command to the watchdog component 1140 (not shown), wherein the at least one operational command comprises at least one command to output at least one warning signal. The sensor error detection component 1150 may be configured to calibrate the at least one sensor and / or disable the at least one sensor.

[0430] The attack detection and counter component 1160 may be configured to detect and / or counter an unauthorized manipulation of the operating component 1200 and / or 2200. The post-incident analysis component 1170 may be configured to analyze at least one incident detected and / or countered by the attack detection and counter component 1160.

[0431] Fig. 4 depicts the monitoring component 1100 and the operating component 1200. More particularly, it depicts the input 1210, the throughput 1225 and the output 1230 of at least one component of the operating component 1200, passed through monitoring algorithm 1112. Monitoring algorithm 1112 would output at least one state relating to the operating component 1200. The state may be a trustworthy state 1113 and / or an untrustworthy state 1114. The monitoring component may be configured to output at least one degree of trustworthiness regarding each state. For example, operating system 1200 may be rated as 80% trustworthy and 20% untrustworthy (confidence interval).

[0432] Additionally and / or alternatively, monitoring algorithm 1112 may output at least one state related to the operating component 1200 such as state 1136, state 1137, state 1138 ... The states may or may not be mutually exclusive. The states may be affected by the constraint component 1130. More particularly, the states may be affected by threshold 1132, threshold 1134 and / or threshold 1135, which are generated according to constraint 1131 and constraint 1133.

[0433] The constraints component 1130 may be configured to comprise at least one constraint wherein the at least one constraint may be utilized to generate at least one threshold. The at least one constraint may comprise but may not be limited to at least one guideline, at least one plan, at least one test plan, at least one directive, at least one norm, at least one mandate, and / or at least one principle.

[0434] Fig. 5 depicts an example for the use of the apparatus according to embodiments of the described invention. The operating component 1200' here implements a visual odometry algorithm 1220' to output a path 1230' for a drone to follow. The operating component 1200' may also comprise the drone. In this example, the drone may be configured as an autonomous delivery drone providing service in a certain area.

[0435] The acquiring component 1300' here would acquire input, throughput and / or output (path) 1230'. The acquiring component may also acquire data that might not specifically be considered input, throughput and / or output. For example, GPS location may be acquired by the acquiring component 1300'. The acquiring component 1300' may transmit the data acquired to a processing component 1400' and / or the monitoring component 1100'. The processing component 1400' would then transmit the now processed data to the monitoring component 1100', more particularly a monitoring algorithm 1112'. The monitoring algorithm 1112' would then classify the operating module 1200' into a trustworthy 1113' and / or untrustworthy state 1114'. Additionally, the monitoring algorithm 1112' may also classify the operating module as being into a trusted state 1136', a buffer state 1137' or a danger state 1138', wherein the trusted state 1136', the buffer state 1137' and the danger state 1138' are based on the location of the drone and / or the path 1230' of the drone. The monitoring component 1100' may also be configured to detect OD / ODD (Operational Domain / Operational Design Domain) violations, short term and long term variation(s), and / or distribution drift, to quantify output confidence, and to report limit violations.

[0436] While in the above, a preferred embodiment has been described with reference to the accompanying drawings, the skilled person will understand that this embodiment was provided for illustrative purpose only and should by no means be construed to limit the scope of the present invention, which is defined by the claims.

[0437] Whenever a relative term, such as "about", "substantially" or "approximately" is used in this specification, such a term should also be construed to also include the exact term. That is, e.g., "substantially straight" should be construed to also include "(exactly) straight".

[0438] Whenever steps were recited in the above or also in the appended claims, it should be noted that the order in which the steps are recited in this text may be accidental. That is, unless otherwise specified or unless clear to the skilled person, the order in which steps are recited may be accidental. That is, when the present document states, e.g., that a method comprises steps (A) and (B), this does not necessarily mean that step (A) precedes step (B), but it is also possible that step (A) is performed (at least partly) simultaneously with step (B) or that step (B) precedes step (A). Furthermore, when a step (X) is said to precede another step (Z), this does not imply that there is no step between steps (X) and (Z). That is, step (X) preceding step (Z) encompasses the situation that step (X) is performed directly before step (Z), but also the situation that (X) is performed before one or more steps (Yl), ..., followed by step (Z). Corresponding considerations apply when terms like "after" or "before" are used.

[0439] Reference Numbers:

[0440] 1000 Apparatus

[0441] 1100 Monitoring component

[0442] 1110 Set of monitoring algorithms

[0443] 1112 Monitoring algorithm

[0444] 1113 Trustworthy State

[0445] 1114 Untrustworthy State

[0446] 1115 AI-Monitoring component

[0447] 1120 Calibrating component

[0448] 1130 Constraints component

[0449] 1131 Constraint 1

[0450] 1132 Threshold 1

[0451] 1133 Constraint 2

[0452] 1134 Threshold 2

[0453] 1135 Threshold 3

[0454] 1136 State 1

[0455] 1137 State 2

[0456] 1138 State 3

[0457] 1140 Watchdog component

[0458] 1150 Sensor-Error Detection component

[0459] 1160 Attack Detection and Counter component

[0460] 1170 Post-Incident Analysis Component

[0461] 1200 Operating component

[0462] 1210 Input to operation in operating component

[0463] 1220 Operation

[0464] 1225 Throughput of operation in operating component

[0465] 1230 Output of operation in operating component

[0466] 1300 Acquiring component

[0467] 1400 Processing component

[0468] 1500 Logger component

[0469] 1100' Monitoring component in example

[0470] 1112' Monitoring algorithm in example

[0471] 1113' Trustworthy state in example

[0472] 1114' Untrustworthy state in example

[0473] 1136' Trusted state

[0474] 1137' Buffer state

[0475] 1138' Danger state 1200' Operating module in example

[0476] 1220' Visual odometry algorithm

[0477] 1230' Output path

[0478] 1300' Acquiring component in example 1400' Processing component in example

[0479] 2000 External system

[0480] 2200 Operating component (external)

[0481] 2210 Input to operation (external) in operating component (external) 2220 Operation (external)

[0482] 2225 Throughput of operation (external) in operating component (external)

[0483] 2230 Output of operation (external) in operating component (external)

[0484] 3000 External system (other)

Claims

Claims1. An apparatus for monitoring of an external system, wherein the external system comprises an operating component, wherein the apparatus comprises a monitoring component, wherein the monitoring component is configured to analyze the components of the operating component and / or at least one data wherein the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component , wherein the operating component is configured to execute at least one operation wherein the at least one operation is monitored by the monitoring component.

2. The apparatus according to the preceding claim, wherein the apparatus comprises the operating component, wherein the operating component is configured to output at least one operational command wherein the at least one operational command is configured to affect the apparatus and / or external system.

3. The apparatus according to any of the preceding claims wherein the monitoring module is configured to classify at least one component of the operating component and / or at least one component of the apparatus into a trustworthy state and / or an untrustworthy state, and / or into at least one state according to the at least one data and at least one threshold.

4. The apparatus according to any of the preceding claims wherein the monitoring component is configured to improve the operating component, wherein the monitoring component is configured to modify at least one component of the operating component and / or at least one data related to at least one component of the operating component.

5. The apparatus according to any of the preceding claims wherein, wherein the operating component comprises at least one Al-Operating component, wherein the monitoring component comprises at least one Al-monitoring component, wherein the Al-monitoring component is configured to analyze the components of the operating component and / or at least one data wherein the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component by making use of at least one Al algorithm, wherein the Al-operating component and the Al-monitoring component are configured as a dual Al paradigm.

6. The apparatus according to any of the preceding claims, with the features of claim 3, wherein the monitoring component comprises at least one constraintscomponent, wherein the constraints component comprises at least one constraint that the operating component must abide by, wherein the constraints component is configured to generate the at least one threshold according to at least one constraint and / or wherein the at least one state is dependent on the at least one threshold and / or the at least one constraint.

7. The apparatus according to any of the preceding claims, wherein the apparatus comprises an acquiring component, wherein the acquiring component is configured to acquire at least one data and / or at least one metadata related to the at least one data , wherein the acquiring component comprises at least one sensor, wherein the at least one sensor is configured to acquire at least one data and / or at least one metadata related to the at least one data.

8. The apparatus according to any of the preceding claims with the features of claim 7, wherein the acquiring component is configured to acquire at least one data from at least one component of the operating component.

9. The apparatus according to any of the preceding claims with the features of claim 7, wherein the monitoring component comprises a sensor error detection component, wherein the sensor error detection component is configured to detect at least one sensor error.

10. The apparatus according to any of the preceding claims wherein the monitoring component comprises an attack detection and counter component, wherein the attack detection and counter component is configured to detect at least one unauthorized manipulation of any component of the operating component, and / or wherein the attack detection and counter component is configured to counter at least one unauthorized manipulation of any component of the operating system.

11. The apparatus according to any of the preceding claims with the features of claim 10, wherein the monitoring component comprises a post-incident analysis component, wherein the post-incident analysis component is configured to analyze at least one incident detected and / or countered by the attack detection and counter component, wherein the post-incident analysis module is configured to transmit at least one analysis result to an external system and / or the external system.

12. The apparatus according to any of the preceding claims wherein the apparatus comprises a logger component, wherein the logger component is configured tostore at least one data, and / or at least one metadata.

13. A method for monitoring of an external system, wherein the external system is configured to operate an operating component, wherein the method comprises operating a monitoring component, wherein operating the monitoring component comprises analyzing the components of the operating component and / or at least one data wherein operating the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component, wherein operating the operating component comprises executing at least one operation wherein the at least one operation is monitored by the monitoring component.

14. The method according to the preceding claim, wherein the method comprises operating the operating component, wherein operating the operating component comprises outputting at least one operational command wherein the at least one operational command is configured to affect the external system and / or the components operated by the method.

15. The method according to any of the preceding method claims, wherein operating the monitoring component comprises classifying at least one component of the operating component and / or at least one component of the method into a trustworthy state and / or an untrustworthy state, and / or classifying at least one data related to at least one component of the operating component, into at least one state according to the at least one data and at least one threshold.

16. The method according to any of the preceding method claims wherein operating the monitoring component comprises improving the operating component, wherein operating the monitoring component comprises modifying at least one component of the operating component and / or at least one data related to at least one component of the operating component.

17. The method according to any of the preceding method claims wherein, wherein operating the operating component comprises operating at least one Al-Operating component, wherein operating the monitoring component comprises operating at least one Al-monitoring component, wherein operating the Al-monitoring component comprises analyzing the components of the operating component and / or at least one data wherein operating the at least one data is configured as input, throughput, and / or output according to each component comprised in the operating component by making use of at least one Al algorithm, whereinoperating the Al-operating component and operating the Al-monitoring component comprise operating a dual Al paradigm.

18. The method according to any of the preceding method claims, with the features of claim 15, wherein operating the monitoring component comprises operating at least one constraints component, wherein operating the constraints component comprises implementing at least one constraint that the operating component must abide by, wherein operating the constraints component comprises generating the at least one threshold according to at least one constraint and / or wherein the at least one state is dependent on the at least one threshold and / or the at least one constraint.

19. The method according to any of the preceding method claims, wherein the method comprises operating an acquiring component, wherein operating the acquiring component comprises acquiring at least one data and / or at least one metadata related to acquiring the at least one data, wherein operating the acquiring component comprises operating at least one sensor, wherein operating the at least on sensor comprises acquiring at least one data and / or at least one metadata related to the at least one data.

20. The method according to any of the preceding method claims with the features of claim 19, wherein operating the acquiring component comprises acquiring at least one data from at least one component of the operating component.

21. The method according to any of the preceding method claims with the features of claim 19, wherein operating the monitoring component comprises operating a sensor error detection component, wherein operating the sensor error detection component comprises detecting at least one sensor error.

22. The method according to any of the preceding method claims wherein operating the monitoring component comprises operating an attack detection and counter component, wherein operating the attack detection and counter component comprises detecting at least one unauthorized manipulation of any component of the operating component, and / or wherein operating the attack detection and counter component comprises operating counter at least one unauthorized manipulation of any component of the operating component.

23. The method according to any of the preceding method claims with the features of claim 22, operating the monitoring component comprises operating a post-incidentanalysis component, wherein operating the post-incident analysis component comprises analyzing at least one incident detected and / or countered by the attack detection and counter component, wherein operating the post-incident analysis component comprises transmitting at least one analysis result to an external system and / or the external system.

24. The method according to any of the preceding method claims wherein the method comprises operating a logger component, wherein operating the logger component comprises storing at least one data, and / or at least one metadata.

Images