System for recording and analysing flight data

Abstract

The invention relates to a system for recording and analysing flight data of an aircraft (A) flying at low altitude below 3000 meters, the aircraft (A) comprising integrated sensors (A1), the system comprising: - an on-board housing (1) in the aircraft (A), containing independent sensors (11) that are separate from the integrated sensors (A1) and operate autonomously without any connection to the integrated sensors (A1) of the aircraft (A), the on-board housing (1) comprising a connectivity module (14) for transmitting the flight data from the independent sensors (11) to the ground, the connectivity module (14) supporting various communication technologies, in particular 3G and 4G, thus ensuring continuous and reliable transmission, and - an analysis module (2) that remains on the ground, is equipped with reception means (23) and receives the flight data transmitted by the connectivity module (14) of the on-board housing (1), the analysis module (2) comprising analysis means (21) for analysing the received flight data, characterized in that it furthermore comprises: - an automatic network switch (141) for automatically switching between the available networks, giving preference to the most stable and fastest connections, in order to ensure uninterrupted transmission of the data, even in the event of signal loss or weak connection, and - a temporary buffer memory (13) for ensuring continuity of the recordings by managing any connection losses, the temporary buffer memory (13) storing the flight data during periods when the connection is interrupted and, once the connection has been restored, the data stored in the temporary buffer memory (13) are automatically transmitted, ensuring that all flight information is correctly recorded and sent, even during a loss of connection.

Description

flight data recording and analysis system

[0001] The present invention relates to an advanced flight data recording and analysis system for general aviation, and more particularly for aircraft flying at low altitudes, typically below 3,000 meters (10,000 feet). The invention is therefore particularly applicable to small aircraft, such as those found in flying clubs.

[0002] In the field of commercial airliners, which fly long distances at high altitudes (approximately 10,000 meters), there is a conventional recording system known as a "black box." The black box is a data recorder on board an aircraft designed to provide flight parameters and pilot conversations for accident investigation. Flight recorders are protected by a casing that is resistant to shock, fire, and immersion.There are two types of flight recorder: the cockpit voice recorder (CVR), which allows for the playback of conversations as well as audible alarms and, to some extent, suspicious noises in the cockpit; and the flight data recorder (FDR), for the analysis of the position of the control surfaces, the operation of the engines and systems, and other data that may help to determine the origin of a possible accident.

[0003] Black boxes are usually placed at the rear of the aircraft, as this is the part that is generally best preserved during an impact with the ground or sea.

[0004] To retrieve the data recorded by the black box, the aircraft must be on the ground: in-flight retrieval is not possible due to technical limitations. In particular, satellite connectivity does not allow for the widespread transmission of the corresponding data volumes. Generally, an external recording module, called a QAR (Quick Access Recorder), associated with the black box, is used to facilitate the retrieval of recorded data without having to directly access the black box. This QAR module is a copy of the FDR (Flight Data Recorder).

[0005] Therefore, the data from the black box or its external recording module is only used for subsequent analysis, which will be carried out after the aircraft has landed, to reconstruct the flight conditions that could explain an incident or accident.

[0006] The 3,000-meter (10,000-foot) limit marks the boundary between lower and upper airspace. Light aircraft, such as those used in general aviation, frequently fly in lower airspace, i.e., at altitudes below 3,000 meters, for several technical and regulatory reasons, including pressurization and engine type. Regarding pressurization, light aircraft generally lack pressurization systems designed for high-altitude flight because the reduction in atmospheric pressure above 3,000 meters necessitates specialized systems to maintain safe flight conditions for passengers and pilots. As for engine type, light aircraft engines (often piston engines) are designed for optimal performance at lower altitudes.At higher altitudes, air density decreases, which negatively affects the performance of piston engines, unlike the turbine engines used in commercial aircraft.

[0007] These constraints explain why light aircraft are mainly limited to lower airspace, where flight conditions are better suited to their structure and propulsion systems.

[0008] Prior art includes documents US2019005744A1 and EP3616173B1, which describe an onboard unit equipped with internal sensors (accelerometers, gyroscopes, pressure sensors, GPS, etc.). However, there is no evidence to suggest that this unit operates without any connection to existing sensors or avionics systems. On the contrary, in document US2019005744A1, the mention of a "GPS feed or an internal GPS" clearly implies that the system may depend on data streams provided by the aircraft itself. This document focuses primarily on the organization of transmissions (Wi-Fi, VHF, HF, Satcom) and on data processing on the ground, but not on an architecture that is completely autonomous from onboard sensors.

[0009] The systems described in documents US2019005744A1 and EP3616173B1 operate in a completely different context: that of commercial aircraft, often equipped with high-speed satellite or Wi-Fi links designed to transfer large volumes of data. In these environments, the issue is not the stability of the link, but rather the capacity to store and process the transmitted data.

[0010] Conversely, in the case of light aviation, aircraft fly below 3,000 m (10,000 ft), with unstable cell coverage. The main constraint is the variability of signal quality: a typical flight successively crosses areas with strong, weak, or no signal.

[0011] The objective of the proposed system is therefore to ensure the resilience of the data flow in this context, without loss, without temporal interruption and without human intervention.

[0012] This approach is specific to general aviation: it does not concern pressurized aircraft or high-altitude flights (segment covered by documents US2019005744A1 and EP3616173B1), but rather club or training aircraft, not equipped with oxygen on board, operating in the lower flight range.

[0013] This restricted scope reflects a distinct reality of use and constitutes a technical and operational differentiation.

[0014] None of the cited documents describe or suggest such a functional interaction.

[0015] Documents US2019005744A1 and EP3616173B1 mention data collection and transmission, but without linking measured radio quality to acquisition rate or sending policy.

[0016] We are also familiar with documents US2017200379A1 and US2021329738A1. Document US2017200379A1 mentions a change in the means of communication (for example, between 4G and Satcom), but does not describe any change in the cellular network or any link to FIFO memory or flow prioritization. As for document US2021329738A1, it deals with prioritization, but in an abstract way, without linking it to signal quality or data persistence.

[0017] The present invention is therefore in a different context from the black box, since it seeks to collect in flight the data recorded during the flight in order to analyze them immediately on the ground, in order to know in real time the flight conditions of the aircraft and to calculate values ​​useful for the flight.

[0018] The invention aims to improve flight safety and monitoring by offering optimized data management and in-depth analysis of flight conditions. It is distinguished by its ability to autonomously record flight data, transmit this data in real time, and enrich this data on the ground through post-processing.

[0019] To this end, the present invention proposes a system for recording and analyzing flight data from an aircraft (A) flying at low altitude below 3000 meters, this aircraft (A) comprising integrated sensors (A1), the system comprising:

[0020] - an onboard unit (1) in the aircraft (A) containing independent sensors (11), separate from the integrated sensors (A1), and which operate autonomously without connection to the integrated sensors (A1) of the aircraft (A), the onboard unit (1) comprising a connectivity module (14) for transmitting flight data from the independent sensors (11) to the ground, the connectivity module (14) supporting various communication technologies, including 3G and 4G, thus ensuring continuous and reliable transmission, and

[0021] - an analysis module (2), remaining on the ground and equipped with receiving means (23), which receives the flight data transmitted by the connectivity module (14) of the on-board unit (1), the analysis module (2) comprising means for analyzing (21) the received flight data,

[0022] characterized in that it further comprises:

[0023] - an automatic network switch (141) to automatically switch between available networks, prioritizing the most stable and fastest connections, in order to ensure uninterrupted data transmission, even in the event of signal loss or weak connection, and

[0024] - a temporary buffer (13) to ensure continuity of recordings by managing possible connection losses, the temporary buffer (13) storing flight data during periods when the connection is interrupted and once the connection is restored, the data stored in the temporary buffer (13) is automatically transmitted, ensuring that all flight information, even during a loss of connection, is correctly recorded and sent.

[0025] This combination of features reflects the complete operation of the system: a self-contained unit with its own sensors and antennas, connected to the ground via a cellular module capable of measuring radio quality, operating automatic switches, managing persistent FIFO memory and maintaining temporal coherence of transmissions.

[0026] These characteristics are not independent: they form a closed loop where radio quality drives the choice of protocol, the network used, buffer management, and the nature of the transmitted data. This operation guarantees temporal continuity and service resilience despite highly fluctuating radio conditions.

[0027] The advantage provided is significant resilience in environments with highly fluctuating connectivity, particularly in light aviation. The system maintains data traceability and integrity even during prolonged outages, while dynamically adapting its transmission strategy.

[0028] Unlike documents US2019005744A1 and EP3616173B1, the system of the invention is characterized by a truly autonomous unit, which incorporates its own sensors and antennas, excluding its own power supply. This unit does not rely on any signal, bus, or onboard instrument.

[0029] This independence constitutes a new technical response compared to documents US2019005744A1 and EP3616173B1: it allows for rapid installation on any light aircraft, without modification of existing systems or connection to the avionics network.

[0030] It solves a unique problem specific to general aviation: how to equip non-certified or training aircraft with a reliable data recording and transmission device, without intervening on the onboard systems.

[0031] This "standalone" aspect of the case is neither described nor suggested in documents US2019005744A1 and EP3616173B1.

[0032] Advantageously, the connectivity module can constantly monitor the mobile networks available in its environment and measure parameters, which are collected at regular intervals to evaluate the performance of the network to which the module is currently connected and those of other networks within range, and when the connectivity module (14) detects that an available network offers better performance than the one to which it is connected, it automatically switches to that network, relying on optimization algorithms that take into account the following three criteria, namely the stability of the connection, the maximum throughput and the minimization of interruptions.

[0033] Advantageously, the on-board unit can integrate means of transmitting essential data capable of prioritizing the sending of essential information, such as critical flight parameters (altitude, speed, GPS position), before transmitting secondary data, particularly in the event of bandwidth limitation or momentary loss of connection.

[0034] This live transmission capability allows ground teams to monitor flight performance and conditions in real time, facilitating a rapid response in case of problems.

[0035] The system proposed by the invention adopts a fully adaptive behavior: connection quality influences how a 4G connection is used (USSD / SMS / UDP / HTTP), the buffer ensures temporal continuity, and prioritization guarantees the transmission of essential data, even in degraded mode. This integrated regulation does not appear in any prior art.

[0036] Advantageously, the independent sensors can be selected from GPS for position tracking, accelerometers and gyroscopes to measure aircraft movement and orientation, and environmental sensors to record temperature and atmospheric pressure. These sensors are completely independent of the aircraft's systems, facilitating installation in a wide variety of aircraft without requiring complex integration with existing onboard systems. Sensor independence also reduces installation costs and complexity, while making the system adaptable to a broad range of aircraft. This also simplifies maintenance, as the system is not dependent on other aircraft components.

[0037] This approach makes it possible to offer flexible and reliable flight data recording, without being limited by the specific technical configuration of the aircraft.

[0038] One of the major innovations of this invention is its ability to transmit flight data in real time, unlike traditional flight data recorders that store data locally and require manual retrieval after the flight. The connectivity module supports various communication technologies, including 3G and 4G, and even future network standards, thus ensuring continuous and reliable transmission.

[0039] Real-time transmission is enhanced with automatic network switching, enabling automatic switching between available networks, prioritizing the most stable and fastest connections to ensure uninterrupted data transmission, even in the event of signal loss or weak connection.

[0040] Another key feature of this invention is the use of machine learning algorithms to analyze flight data and calculate additional advanced parameters, thereby providing valuable information for flight performance analysis and safety management.

[0041] The system analysis module may, for example, include means of automatic analysis of flight data, using machine learning algorithms capable of analyzing recorded data to identify specific patterns, enabling the automatic calculation of parameters, such as aircraft drift, which can be influenced by weather conditions, and the evaluation of necessary corrections in real time.

[0042] The system analysis module may also include means for automatically determining flight phases capable of identifying the different phases of flight (climb, cruise, descent, landing) without human intervention, for precise monitoring of aircraft performance and a better understanding of flight conditions in real time.

[0043] The analysis module may also include means of predicting potential incidents using machine learning algorithms that analyze real-time data and detect anomalies, to anticipate risky situations or potential incidents (such as unexpected turbulence or trajectory deviations) and alert pilots or ground crews to take preventive action.

[0044] Advantageously, the analysis module also includes means of continuous improvement of algorithm performance via machine learning, also using the collected data to progressively improve algorithm performance, allowing for better optimization of flight safety and efficiency.

[0045] This automatic calculation of advanced parameters using machine learning improves the understanding of flight conditions and contributes to better anticipation of incidents, thus offering increased safety for general aviation.

[0046] The invention is based on a control loop that combines continuous radio quality measurement, automatic network failover, persistent FIFO management, and data prioritization to ensure temporal continuity and the integrity of transmitted information. The proposed system is designed for a distinct operational context: that of light aviation using cellular networks subject to significant variations in coverage.

[0047] Documents US2019005744A1 and EP3616173B1 do not demonstrate the complete independence of the sensors from the avionics systems. Neither document describes or suggests such a functional interaction between these various elements.

[0048] The invention will now be described in greater detail, with reference to the accompanying drawings, giving by way of non-limiting example, one embodiment of the invention.

[0049] In the figures:

[0050] This is a representational view of an aircraft equipped with the on-board unit of the invention.

[0051] This is a very schematic view of a basic version of the embedded unit of the invention,

[0052] This is a very schematic view of a basic version of an analysis module for the invention.

[0053] This is a schematic view of a more advanced version of the embedded housing of the invention,

[0054] This is a schematic view of a more advanced version of the analysis module of the invention,

[0055] This is a schematic view illustrating the calculation of the aircraft's drift using the invention.

[0056] It illustrates the different phases of taxiing and flight of an aircraft, and

[0057] This illustrates the detection of a hard landing.

[0058] The proposed flight data recording and analysis system is designed to operate autonomously in aircraft A, independently of existing integrated sensors A1. This system consists of two main components: a unit 1 installed in aircraft A and an analysis module 2, which remains on the ground.

[0059] The unit 1 is responsible for recording flight data. It contains various autonomous sensors, such as those for altitude measurement, GPS position and other critical parameters.

[0060] Analysis Module 2 receives data transmitted by the unit 1 onboard aircraft A. It is responsible for analyzing the recorded flight data, allowing users to extract relevant information for flight performance and safety assessments. The ground-based Analysis Module 2 may also include user interfaces for viewing data in real time and generating detailed reports based on the analyses performed.

[0061] This architecture allows for accurate and secure data collection, ensuring that all essential information is recorded and accessible for in-depth analysis after each flight.

[0062] The unit 1 is designed for installation in the aircraft without requiring modifications to existing systems. It operates independently, powered by an external source (via USB), and uses its own independent sensors. Installation is simple, requiring only a connection to a 5V power supply via a standard USB port.

[0063] A basic version for the on-board unit 1 includes independent sensors 11, a recording module 12, a temporary buffer 13, a connectivity module 14 and a power supply 15 for the unit 1. This basic version is shown on the.

[0064] The core of unit 1 is a microcomputer, such as a Raspberry Pi, capable of processing data from the sensors. The Raspberry Pi was chosen for its compact size, low power consumption, and sufficient processing power to handle the acquisition and storage of flight data. Unit 1 can also use a compatible add-on module, such as a SenseHat, to add specific measurement functionalities.

[0065] The compatible add-on module 11 is equipped with several independent sensors, including a GPS sensor 111, accelerometers 112, gyroscopes 113, and environmental sensors 114, which enable detailed monitoring of environmental parameters and aircraft A's movements. These sensors are used to record data such as altitude, rotation speed, aircraft bank angle, and pressure variations. This is represented on the...

[0066] The unit 1 is designed not to interfere with the aircraft's integrated A1 sensors. It operates independently, without needing to connect to the integrated A1 sensors. This independence ensures the reliability of the data collected while avoiding potential conflicts with the aircraft's critical systems.

[0067] Connectivity Module 14 supports various communication technologies, such as 3G and 4G, ensuring continuous and reliable transmission of flight data. This connectivity module is implemented via a 4G USB dongle, connected directly to the Raspberry Pi. The 4G USB dongle allows the device to establish a mobile network connection and transmit the collected data in real time to a remote server or monitoring interface. The USB dongle plugs directly into a USB port on the Raspberry Pi, without the need for additional components or complex cabling.

[0068] Thanks to this integration, unit 1 is able to maintain constant communication, even when aircraft A is operating in isolated areas or at high altitudes, where network coverage may be limited. By combining 3G and 4G technologies, it ensures optimal redundancy and connection stability, minimizing data loss during transmission.

[0069] Connectivity modules, such as the Huawei E3372 dongle, feature an automatic failover mechanism between available networks (3G, 4G, LTE) to maintain a stable and optimal connection. This process is managed by the network management module integrated into the device, which continuously assesses the signal quality of available networks and performs dynamic switching based on specific criteria.

[0070] Connectivity module 14 constantly monitors the mobile networks available in its environment. It measures parameters such as:

[0071] • Signal strength (RSSI - Received Signal Strength Indicator),

[0072] • Signal quality (SINR - Signal to Interference plus Noise Ratio),

[0073] • The type of network (3G, 4G, LTE),

[0074] • Latency and available bandwidth.

[0075] This information is collected at regular intervals to assess the performance of the network to which the module is currently connected and that of other networks within range.

[0076] When connectivity module 14 detects that an available network offers better performance than the one it is currently connected to, it automatically switches to that network. This switchover is based on optimization algorithms that take into account the following criteria:

[0077] • Connection stability: The module prioritizes the most stable connections to avoid data loss, even if the throughput is not the highest.

[0078] • Maximum throughput: Depending on the signal quality, the module may decide to switch to a network offering better download and upload speeds.

[0079] • Minimizing interruptions: The transition between networks is seamless, without significant interruption of the data flow, thanks to continuity management protocols (such as LTE handover).

[0080] Connectivity Module 14 is designed to prioritize the fastest available network. By default, it will attempt to connect to a 4G LTE network. However, if the coverage or quality of this network becomes insufficient, it will automatically switch to 3G or a lower-bandwidth LTE network, thus ensuring a connection, even if it is slower.

[0081] If the connectivity module 14 finds a stronger 4G network during continuous monitoring, it will automatically switch back to it. This flexibility is crucial in environments where mobile coverage can fluctuate, such as during high-altitude flights or in rural areas.

[0082] The unit 1 incorporates a temporary buffer 13 to ensure the continuity of flight recordings by managing any potential connection losses. This buffer 13 is essential for storing flight data during periods when the connection is interrupted.

[0083] Buffer 13 is implemented as a FIFO (First In First Out) structure. In this configuration, flight data is added to the end of the queue and retrieved in the order it was added. This method ensures that records are processed and sent in chronological order, thus preserving the sequence of flight events.

[0084] The data stored in buffer 13 is serialized before being written to the Raspberry Pi's SD card. Serialization converts data objects into a binary or text file format that can be easily stored and retrieved. This simplifies memory management and optimizes storage space on the SD card. One example of a serializable format is JSON. The serialized data is stored on the Raspberry Pi's SD card, providing a non-volatile storage solution. In the event of a power outage, the data is preserved on the SD card, ensuring that no flight information is lost.

[0085] When new flight data is collected by the sensors, it is immediately added to the end of the FIFO queue. If the connection is active, the data is also transmitted in real time. In the event of a connection loss, the data continues to be stored in buffer 13 until its maximum capacity is reached. Once the connection is restored, unit 1 begins retrieving the data from FIFO buffer 13 and serializing it again for transmission to the server or monitoring interface. This process ensures that all data collected during the interruption is sent in order, thus guaranteeing the integrity of the flight records.

[0086] If buffer 13 reaches a certain capacity threshold, unit 1 can reduce the data sampling rate. This means the sensors will record flight data less frequently, thus decreasing the data flow into buffer 13. For example, instead of taking measurements every second, unit 1 might switch to taking measurements every two or three seconds.

[0087] The unit 1 may also include means of securing transmissions 142, guaranteeing the protection of flight data when it is sent to the ground unit. For this purpose, it uses the HTTPS (HyperText Transfer Protocol Secure) protocol to make secure requests.

[0088] The first box implements HTTPS to encrypt the data exchanged between the Raspberry Pi's 4G dongle and the ground server. This communication protocol combines HTTP with the security of SSL / TLS. The first box uses a suitable library (e.g., requests for Python) that supports HTTPS. This library facilitates the creation of secure requests to the server. The ground server must be configured with valid SSL / TLS certificates to establish secure HTTPS connections. This includes managing certificates on the Raspberry Pi to allow verification of the server's authenticity.

[0089] The unit 1 can also include means for prioritizing essential data 143, allowing the transmission of critical flight information to be prioritized. This feature is crucial to ensuring that the most important data is transmitted first, even in the event of bandwidth limitations or connection loss.

[0090] Unit 1 identifies and classifies flight data into two main categories:

[0091] • Essential data

[0092] This information includes critical flight parameters such as:

[0093] o Altitude: Measurement of the aircraft's current altitude, essential for flight safety.

[0094] o GPS position: Geographic coordinates allowing the aircraft's trajectory to be tracked.

[0095] This data is transmitted as a priority, ensuring that it is available in real time for effective monitoring.

[0096] • Less critical data:

[0097] This information includes parameters such as:

[0098] o Speeds: Indicates the speed of the aircraft.

[0099] o Accelerations: Measurement of speed variations on the three axes of the gyroscope

[0100] o Temperature: Internal temperature of the aircraft.

[0101] o Barometric pressure: Indicator of atmospheric conditions.

[0102] This data can be sent after essential information, allowing for efficient bandwidth management. Device 1 uses queues to organize the data to be transmitted. Essential data is placed at the front of the queue, while less critical data is added afterward. When the connection is active, the data is transmitted in the established priority order. In case of bandwidth limitations, the device can adjust the data flow, focusing on transmitting critical information while delaying the transmission of less important data.

[0103] For all signals, using the example below of accelerometer 112, unit 1 implements dynamic sampling to optimize signal recording. This mechanism works as follows:

[0104] • Detection of significant variations:

[0105] Unit 1 continuously monitors accelerometer data. When it detects sufficient variations (for example, significant acceleration or deceleration), it activates high-frequency sampling. This allows for the capture of precise details of aircraft A's movements during critical maneuvers.

[0106] • Normal sampling in the absence of variations:

[0107] When the accelerometers 112 do not detect significant variations, the housing 1 reduces the sampling frequency, which saves memory and reduces the flow of data to be transmitted.

[0108] This dynamic sampling mechanism, combined with data prioritization, optimizes the transmission of essential data and ensures that all critical information is captured and sent without delay, while minimizing the clutter of less critical data.

[0109] The analysis module 2, which remains on the ground, includes receiving means 23 capable of cooperating with the connectivity module 14 of the onboard unit 1, as well as storage means 24. This is visible in Figures 3 and 5. The analysis module 2 mainly includes means for the automatic analysis of flight data 21, collected by the onboard unit 1. This analysis makes it possible to extract significant information on flight performance and safety.

[0110] An example of automatic flight data analysis is the calculation of aircraft drift. Drift is the difference between the aircraft's actual heading and its recorded heading, and it can be influenced by factors such as wind. Let's illustrate the calculation of this drift.

[0111] The system begins by collecting GPS position data P1, P2, P3, P4, and P5 from aircraft A. This raw data may contain noise, requiring processing to obtain reliable results. To reduce this noise, the system applies a digital smoothing method. A very basic smoothing method is the moving average method, where the smoothed value at a given time is calculated by averaging the GPS position values ​​P1 through P5 over a defined time interval. For example, if we take the GPS positions of the last 5 seconds, the smoothed position is given by:

[0112] Smoothed position = (P1+P2+P3+P4+P5) / 5

[0113] where P1 to P5 represent the GPS positions at different times.

[0114] We can thus determine the estimate of the actual trajectory Tr of aircraft A.

[0115] Similarly, the data from gyroscope 113, which measures the aircraft's heading, is also smoothed using the same moving average method. This results in a stable measurement of the aircraft's heading Cm.

[0116] Once the two sets of data (smoothed GPS and smoothed gyroscope) are obtained, the system compares the heading deduced from the smoothed GPS data with that recorded by the gyroscope. The difference between these two values ​​represents the drift D of aircraft A.

[0117] To refine the analysis, the system can correlate the calculated drift D with meteorological data retrieved from METARs (Meteorological Aerodrome Reports). A METAR is a standardized meteorological report that provides information on weather conditions at a specific airport at a given time, including elements such as wind speed and direction. Using this information, the system can determine if the observed drift is consistent with the weather conditions, which can help adjust flight strategies and improve safety.

[0118] The automatic flight data analysis means 21 may include automatic flight phase determination means 211. These means 211 use the altitude data recorded by the onboard unit 1. These data, once smoothed to reduce noise, provide a clear representation of the evolution of the altitude of aircraft A over time.

[0119] The analysis is based on altitude variations over time: Taxiing at departure: The altitude is generally constant and close to zero, as the aircraft moves along the tarmac. Takeoff: A rapid change in altitude is detected, indicating that the aircraft is climbing rapidly. This occurs after reaching sufficient takeoff speed. Cruise: After takeoff, the aircraft reaches a stable cruising altitude. During this phase, altitude data shows little or no significant variation, as the aircraft maintains a constant altitude. Landing: A gradual decrease in altitude is observed, signaling that the aircraft is beginning its descent. Taxiing at arrival: Once the aircraft has touched down, the altitude becomes stable and close to zero, indicating that the aircraft is taxiing again on the tarmac after landing.

[0120] These five phases are illustrated on the.

[0121] Using an algorithm based on altitude and time variation thresholds, the 211 system automatically determines which flight phase is in progress at each instant within the predetermined sequence. For example:

[0122] • If the altitude remains constant for a given period, the system identifies the taxiing phase.

[0123] • If a rapid increase in altitude is detected, it signals the start of the takeoff phase.

[0124] • Once the altitude stabilizes at a certain high value, the system recognizes the cruise phase.

[0125] • The detection of a slow and regular descent corresponds to landing.

[0126] • Finally, when the altitude reaches zero again, the system identifies the taxiing phase upon arrival.

[0127] This description highlights how the analysis module 2 automatically determines flight phases based on their sequence, altitude variations, and time.

[0128] Automatic flight data analysis means 21 may also include means for predicting potential incidents 212, which make it possible to assess landing conditions and identify risks, such as that of a hard landing, which is illustrated in the figure. A hard landing can cause damage to the aircraft structure and compromise passenger safety.

[0129] Means 212 use the data provided by the gyroscope 113 to determine the approach slope of aircraft A. This slope is essential to evaluate the effective trajectory Te of aircraft A, during landing.

[0130] From the approach slope information, the means 212 project an effective straight trajectory Te onto the horizontal axis of runway P. This makes it possible to define a touchdown point Pa on runway P, where aircraft A would touch down without making a flare.

[0131] The concept of flare refers to the curve that aircraft A describes during its final approach. A shallow flare means that the aircraft followed a nearly direct path to the runway, minimizing sudden changes in angle that can lead to a hard landing. Conversely, a large flare could mean that the aircraft performed maneuvers that increase the risk of a hard impact.

[0132] Once the touchdown point Pa is determined, the system measures the distance between this point Pa and the actual touchdown point Pe (the point where the aircraft touches down). This calculation is crucial for assessing the risk of a hard landing.

[0133] • Short distance: If the distance between the actual touch point Pe and the touchdown point Pa is short, this means that aircraft A approached the runway P in a very straight line, which can lead to a hard landing.

[0134] • High distance: Conversely, if this distance is significant, it indicates a significant gap between the approach and the desired landing point, suggesting a more controlled landing.

[0135] To enhance the assessment of the risk of a hard landing, means 212 correlate the measurement of the distance between the point of touch Pe and the point of touchdown Pa with the values ​​of the vertical acceleration signal collected over the same period.

[0136] By analyzing the vertical acceleration signal, the 212 system can determine if the acceleration at touchdown exceeds a predefined threshold, which would indicate a hard landing. Indeed, a high vertical acceleration at impact can be a direct indicator of a harder-than-expected landing.

[0137] If the distance between the destination point and the actual touchdown point is deemed critical and the vertical acceleration at the time of landing exceeds safety thresholds, the system can generate alerts to inform pilots or ground crews of the risks of a hard landing, thus enabling them to take appropriate measures to avoid incidents.

[0138] The flight analysis module 2 can also include means of continuous improvement of algorithm performance through machine learning. This process allows for the analysis of data collected over time and the progressive refinement of anomaly detection algorithms, thereby increasing their accuracy and efficiency, such as in the case of anomaly detection during landing.

[0139] Each aerodrome has aeronautical documentation that specifies the recommended approach slope for each runway centerline, usually set at around 3°. This reference value is crucial for assessing whether the landing will take place under normal conditions.

[0140] During each flight, the system calculates the effective approach gradient using the position and altitude data provided by the onboard unit. This gradient is then compared to the reference value of 3°.

[0141] If the difference between the actual approach gradient and the reference value exceeds a predefined threshold, the system identifies this as an anomaly. For example, an approach gradient of 5° could indicate an excessive approach angle, which could lead to safety risks.

[0142] To improve anomaly detection and account for the specific characteristics of each aircraft type, the system uses machine learning to adjust the standard approach glide slope. This is done by analyzing historical data from previous flights.

[0143] • Calculation of the approach slope for each flight: For each recorded flight, the system calculates its effective approach slope.

[0144] • Determination of a standard slope: By aggregating the approach slopes of all similar flights (for example, using the median), the system establishes a standard approach slope for each type of aircraft.

[0145] The calculated standard approach glide slope is then stored in the ground system's memory. This new reference value is used for future analyses to improve the accuracy of anomaly detection. Over time, this process allows for dynamic adjustment of the reference values ​​based on the actual aircraft performance during landing.

[0146] The results of the analysis are used to further refine the machine learning algorithms. If anomalies are detected repeatedly for a specific aircraft type, the system can adjust the learning parameters to better account for the performance variations of that aircraft model.

[0147] Detected anomalies and adjustments to the standard approach slope are also recorded in reports, enabling operators to identify trends and optimize flight procedures accordingly.

[0148] The internet-connected receiver module 23 acts as the entry point for all data transmitted by the aircraft's onboard unit. This internet connection enables the reception of real-time flight information, leveraging the network infrastructure to ensure fast and reliable transmission. This module is designed to maintain a constant connection, thus allowing a continuous flow of information, even in conditions where network coverage may fluctuate. By being directly connected to the internet, it ensures a reliable link between the flight equipment and the ground control center.

[0149] To ensure the security of data transmitted by the aircraft, the 231 transmission security system uses HTTPS (Hypertext Transfer Protocol Secure) and SSL (Secure Sockets Layer) protocols. HTTPS encrypts communications between the onboard unit and the ground receiving server, guaranteeing that data cannot be intercepted by third parties during its transit over the internet. SSL, by securing the connection, ensures the authentication of the parties and the confidentiality of the exchanged data. This configuration prevents the risk of falsification or interception of transmitted data, thus strengthening the overall security of the transmission system.

[0150] The means of receiving 232 data are implemented as a RESTful web service, hosted on an application server. This web service is capable of processing incoming HTTP requests sent by the onboard device, which transmits the data as POST requests in JSON format. By adhering to REST standards, the server ensures structured and flexible communication, enabling real-time reception of flight information while guaranteeing the integrity of the received data. This system is designed to seamlessly manage data exchange, thus facilitating integration with tracking and analysis applications without any loss of information.

[0151] The 24-hour storage system is designed to store all flight data received by the 23-hour receiving module. This system uses a database, such as MySQL, to archive information in a structured and reliable manner. The database allows the received information to be organized into tables and ensures quick access for analysis or auditing purposes. Thanks to MySQL, the 24-hour storage system is capable of managing large volumes of data while guaranteeing its integrity, availability, and security, thus meeting the stringent requirements for tracking and storing flight information.

[0152] It can be said that the system's operation relies primarily on a real-time regulation loop structured around three main mechanisms:

[0153] 1. Continuous radio quality measurement: The cellular module continuously evaluates indicators such as RSSI, SINR, latency, and throughput. Unlike existing systems that rely on only one type of connection (Wi-Fi, satellite, etc.), this one utilizes several modes of 4G network usage, depending on signal quality:

[0154] • USSD, used when connectivity is minimal. This mode does not require a data session and allows direct exchange via the network signaling channel. It guarantees basic communication even when the network does not support IP traffic.

[0155] • SMS, used to transmit small data frames when the cellular network is too weak for a TCP / IP session, but stable for short message exchanges. SMS offers a good compromise between reliability and latency.

[0156] • UDP, used for sending lightweight packets in real time when the connection is stable but the bandwidth is limited. This mode does not require formal acknowledgment, but minimizes latency for low-criticality data.

[0157] • HTTP, used when the connection is good enough for a complete transfer. This protocol ensures reliable acknowledgment and retransmission of packets, thus guaranteeing the completeness of the data sent.

[0158] 2. Automatic network failover: When a degradation is detected, the device automatically switches to the most stable available network, without external intervention. This failover is based on a combined evaluation of the preceding indicators (RSSI, SINR, latency, throughput) and guarantees continuity of the communication link.

[0159] 3. Persistent FIFO buffer: data is recorded chronologically. In case of interruption, it is stored in a persistent FIFO queue and then retransmitted in the same order as soon as the signal returns, ensuring temporal coherence and completeness of the stream.

[0160] Prioritizing essential data frames is advantageous: the system distinguishes critical data (position, altitude, vital parameters) from secondary information (vibrations, environment). In case of reduced bandwidth, only essential data continues to be transmitted, ensuring minimal but complete monitoring of the ongoing flight.

[0161] It should be noted that the scope of the invention may be exclusively limited to general aviation, i.e., aircraft flying below 3,000 m (10,000 ft) and unpressurized, in order to distinguish it from onboard systems intended for commercial or long-haul aircraft, such as those cited in documents US2019005744A1 and EP3616173B1. The latter are designed for a completely different environment (high-altitude flights, stable satellite connections, integrated power supply), whereas the invention here applies to autonomous, unconnected devices subject to varying radio conditions.

[0162] List of figure references:

[0163] A Plane

[0164] A1 Integrated Sensors

[0165] 1 On-board unit

[0166] 11 Independent Sensors

[0167] 111 GPS Sensor

[0168] 112 Accelerometer

[0169] 113 Gyroscope

[0170] 114 Environmental Sensors

[0171] 12 Recording Module

[0172] 13 Temporary buffer memory

[0173] 14 Connectivity Module

[0174] 141 Automatic Network Switch

[0175] 142 Means of securing transmissions

[0176] 143 Data Prioritization Method

[0177] 15 Power supply for the enclosure

[0178] 16 Microcomputer

[0179] 2 Ground Analysis Module

[0180] 21 Methods of analyzing flight data

[0181] 211 Means of determining flight phases

[0182] 212 Means of predicting potential incidents

[0183] 22 Means of continuous improvement

[0184] 23 Reception Module

[0185] 231 Means of securing transmissions

[0186] 232 Means of receiving data

[0187] 24 Storage System

[0188] P1 – P5 Recorded GPS Coordinates

[0189] Tr Estimation of the actual trajectory

[0190] Cm Magnetic Cap recorded

[0191] D Drift

[0192] P Track

[0193] The aircraft's effective trajectory

[0194] Point of arrival

[0195] Effective touch point

Claims

Flight data recording and analysis system for an aircraft (A) flying at low altitude below 3000 meters, this aircraft (A) comprising integrated sensors (A1), the system comprising: an onboard unit (1) in the aircraft (A) containing independent sensors (11), separate from the integrated sensors (A1), and which operate autonomously without connection to the integrated sensors (A1) of the aircraft (A), the onboard unit (1) comprising a connectivity module (14) for transmitting flight data from the independent sensors (11) to the ground, the connectivity module (14) supporting various communication technologies, including 3G and 4G, thus ensuring continuous and reliable transmission, and an analysis module (2), remaining on the ground and equipped with receiving means (23), which receives the flight data transmitted by the connectivity module (14) of the onboard unit (1),the analysis module (2) comprising means for analyzing received flight data, characterized in that it further comprises: an automatic network switch (141) for automatically switching between available networks, favoring the most stable and fastest connections, in order to ensure uninterrupted data transmission, even in the event of signal loss or weak connection, and a temporary buffer (13) to ensure continuity of recordings by managing any connection losses, the temporary buffer (13) storing flight data during periods when the connection is interrupted and once the connection is restored, the data stored in the temporary buffer (13) is automatically transmitted, ensuring that all flight information, even during a loss of connection, is correctly recorded and sent. System according to claim 1, wherein the connectivity module (14) constantly monitors the mobile networks available in its environment and measures parameters, which are collected at regular intervals to evaluate the performance of the network to which the module is currently connected and those of other networks within range, and when the connectivity module (14) detects that an available network offers better performance than the one to which it is connected, it automatically switches to that network, relying on optimization algorithms that take into account the following three criteria, namely connection stability, maximum throughput and minimization of interruptions. System according to claim 1, wherein the on-board unit (1) further comprises means for transmitting essential data (143) capable of prioritizing the sending of essential information, such as critical flight parameters (altitude, speed, GPS position), before transmitting secondary data, particularly in the event of bandwidth limitation or momentary loss of connection. System according to any one of the preceding claims, wherein the independent sensors (11) of the on-board unit (1) are selected from GPS (111) for position tracking, accelerometers (112) and gyroscopes (113) for measuring the movements and orientation of the aircraft (A), environmental sensors (114) for recording temperature and atmospheric pressure. System according to any one of the preceding claims, wherein the on-board unit (1) further includes means for securing transmissions (142) to protect the transmitted data by robust security protocols, ensuring that critical flight information remains confidential and protected against unauthorized access. System according to any one of the preceding claims, wherein the flight data analysis means (21), using machine learning algorithms capable of analyzing the recorded data, to identify specific patterns, enabling the automatic calculation of parameters, such as aircraft drift, which may be influenced by weather conditions, and the evaluation of necessary corrections in real time. System according to claim 6, wherein the analysis means (21) include means for automatic determination (211) of flight phases capable of identifying the different phases of flight (climb, cruise, descent, landing) without human intervention, for precise monitoring of aircraft performance and better understanding of flight conditions in real time. System according to claim 6 or 7, wherein the analysis means (21) include means for predicting potential incidents using machine learning algorithms that analyze real-time data and detect anomalies, to anticipate risky situations or potential incidents (such as unexpected turbulence or trajectory deviations) and alert pilots or ground crews to take preventive measures. System according to claim 6, 7 or 8, wherein the analysis module (2) further comprises means for continuous improvement (22) of the performance of the algorithms via machine learning also using the data collected to progressively improve the performance of the algorithms, allowing for better optimization of flight safety and efficiency.

Images