System and method for enhanced asset tracking accuracy and efficiency

Abstract

Provided are systems and methods for optimizing asset tracking devices with reduced data and battery consumption, while improving the efficiency and accuracy of tracking the assets. This includes determining a predicted accuracy of the predicted path segment based on a weight associated with the second sensor and the predicted path segment; determining the predicted accuracy of the predicted path segment exceeds an accuracy threshold; receiving, from the first sensor, updated location data corresponding to the asset; and transmitting, from a network device of the asset tracking device in communication with the processor to a remote device, the location data, the predicted path segment, the predicted accuracy of the predicted path segment, and the updated location data.

Description

[0001] TITLE: SYSTEM AND METHOD FOR ENHANCED ASSET TRACKING ACCURACY AND EFFICIENCY

[0002] Cross-Reference to Related Application

[0003] [1] This application claims the priority of United States Provisional Patent Application No. 63,663,535 filed June 24, 2024 the entirety of which is incorporated herein by reference.

[0004] Field

[0005] [2] The described embodiments relate to tracking for assets such as vehicles, and specifically to improved asset tracking devices with reduced data and battery consumption that provide improved efficiency and accurate tracking of the assets.

[0006] Introduction

[0007] [3] The following is not an admission that anything discussed below is part of the prior art or part of the common general knowledge of a person skilled in the art.

[0008] [4] Asset tracking systems are widely used in modern logistics and fleet management, providing insights into the location, movement, and status of assets such as freight vehicles, ships, and trains. The asset tracking systems leverage a combination of hardware components, such as GPS modules and motion sensors, and software algorithms to track and manage assets in real-time. However, traditional asset tracking systems face challenges related to high battery consumption and data usage, and inconsistent tracking accuracy.

[0009] [5] One of the common challenges faced by traditional asset tracking systems includes high battery consumption. The systems frequently rely on GPS fixes, sensor measurements, and wireless network communication, all of which may drain the battery life of tracking devices rapidly. The high-power consumption results in frequent battery replacements or recharges, increasing maintenance costs and potentially leading to periods of downtime when devices are offline for charging. Secondly, the asset tracking systems often generate large volumes of data that need to be transmitted over telecom or local wireless networks. The high data usage may strain network resources and increase operational costs, making it less sustainable and more expensive to maintain these systems, especially in large-scale deployments.

[0010] [6] Accuracy and reliability of tracking data are major concerns. Inaccurate measurements of device trips, odometer readings, engine hours, or telemetry data may lead to errors in asset tracking, reduced visibility into asset movements, and potential issues with billing or predictive maintenance. These inaccuracies can undermine the effectiveness of tracking systems and result in operational inefficiencies.

[0011] [7] Additional challenges relate to network scalability and energy consumption. High data traffic on wireless networks and cloud data centers may reduce scalability and increase energy consumption, impacting the overall performance and sustainability of the asset tracking systems. The high data traffic may be particularly problematic in environments with a large number of tracked assets, where network congestion and high energy use may degrade system performance.

[0012] [8] Several existing solutions aim to address these challenges in asset tracking systems. One approach involves the development of low-power GPS technologies designed to reduce energy consumption while maintaining accuracy. This is accomplished by decreasing the frequency of GPS data collection. The use of GPS antennae and data collection is a very energy intensive operation for battery powered devices. However, the reduction of data collection frequency incurs a resulting loss of accuracy in the accuracy of the data collected.

[0013] [9] The low-power GPS technologies focus on optimizing GPS performance to extend battery life. However, low-power GPS technologies often still require frequent location fixes, which may limit their effectiveness in reducing battery consumption.

[0014]

[0010] Efficient data compression and transmission protocols are among another possible solutions. The protocols operate to minimize network traffic and data usage by compressing data before transmission and using improved communication methods to ensure that only essential data is sent. The protocols may help reduce operational costs and improve network efficiency. Despite the benefits, the data compression and transmission protocols may lead to delays in data transmission and may require computational resources to implement effectively.

[0015]

[0011] Machine learning and predictive analytics techniques may be employed to improve tracking accuracy. However, the complexity and resource requirements of machine learning models may pose challenges for implementation on low-power devices. Additionally, edge computing and distributed processing architectures are being explored to reduce latency and conserve energy. By processing data closer to the source, these architectures may minimize the amount of data that needs to be transmitted to central servers, reducing network traffic and energy consumption. However, edge computing solutions may be expensive to implement and maintain, and they may require sophisticated infrastructure and expertise.

[0016]

[0012] Further, there is research on dead reckoning methods to supplement GPS data and improve tracking accuracy. Dead reckoning involves using sensor data, such as accelerometers and gyroscopes, to supplement GPS data. The method may provide accurate location data in environments where GPS signals are weak or unavailable.

[0017]

[0013] Some solutions involve algorithms such as Ramer-Douglas-Peucker algorithm for attaching tracking data to specific streets for reducing the number of data points delivered via the network, thereby addressing challenges such as trailer trip and odometer tracking. However, these algorithms can be computationally intensive and may not perform well in areas with poor map data or rapidly changing environments.

[0018]

[0014] Despite these advancements, there remains a need for a comprehensive solution that improves battery and data consumption while enhancing accuracy, reliability, and sustainability in asset tracking systems. For at least these reasons, there exists a need for an improved asset tracking system that enhances the accuracy and efficiency of tracking.

[0019] Summary

[0020]

[0015] The following introduction is provided to introduce the reader to the more detailed discussion to follow. The introduction is not intended to limit or define any claimed or as yet unclaimed invention. One or more inventions may reside in any combination or sub-combination of the elements or process steps disclosed in any part of this document including its claims and figures.

[0021]

[0016] In accordance with one aspect of this disclosure, which may be used alone or in combination with any other aspect, there are provided systems, methods, and computer-readable media for tracking an asset, which improve the accuracy of trip, odometer, and telemetry measurements while reducing the required connectivity data and battery consumption. The disclosed systems, methods and computer-readable media may improve the efficiency and reliability of asset tracking devices, making them more sustainable and cost-effective.

[0022]

[0017] In an aspect of this disclosure, the accuracy and power efficiency can be improved by dynamically adjusting the weights of sensor data and incorporating predictive path models to reduce the frequency of GPS events transmitted to a server. The sensor data may include heading, speed, GPS accuracy, and the rates of change of these data, as well as any signal measurement from a nearby device. Using the sensor data sets, the system may decide that an accurate GPS request is required to calibrate the correct predicted path. Motion sensors can be used to predict the path using historical data including previous GPS, heading, speed, and accelerometer data, thereby delaying GPS and telecommunication events. Additionally, the weight function can be computed based on the accumulated error in previous sensor measurements, such as previous motion and GPS data. During the predicted path correction process, the system can cache the measured sensor data, including both high- and low-quality GPS fixes. The cached data can then be fed, either at once or in batches, to an asset tracking service or on-board geographic information system (GIS) server to accurately regenerate the predicted path taken by the on-board device and derive an accurate measurement of the trip and traveled mileage. Server support can be used to recompute weights considering GIS information and potential deviations from the expected path. Using the generated weights for the predicted location in combination with available GIS information can enable the system to predict which route the device is moving on and project each location that falls below a threshold of the overall weight of the predicted location. The systems and methods of the present disclosure enables the device to operate in low-power mode and report back only when deviating from the received path or confirming adherence to it, minimizing unnecessary communication with the server. This may also reduce the number of events required to predict an accurate location and odometer reading of the device, leading to less battery consumption and data usage.

[0023]

[0018] The systems, methods, and computer-readable media may improve asset tracking by reducing battery consumption and data usage, enhancing accuracy, and promoting sustainability. Further, location accuracy and trip mileage calculation may be improved without the need for computationally intensive filtering algorithms. By dynamically adjusting sensor data weights and incorporating predictive path models, the frequency of GPS updates can be reduced and sensor data usage improved. The reduction may lead to lower operational costs and extended device lifespan.

[0024] Additionally, server-side recalibration and efficient communication protocols can ensure reliable and accurate tracking, reducing errors and improving overall system performance. The systems and methods described herein can address the limitations of existing asset tracking systems, providing a more energy efficient, accurate, and sustainable approach to asset management.

[0025]

[0019] In a first aspect, there is provided a computer-implemented method for tracking an asset using an asset tracking device, comprising: providing, at a memory of the asset tracking device, a location prediction model; receiving, at a processor of the asset tracking device in communication with the memory from a first sensor positioned at the asset, initial location data corresponding to the asset; receiving, at the processor from a second sensor positioned at the asset, sensor data corresponding to the asset; selecting, at the processor, the second sensor as an active sensor for tracking the asset; determining, at the processor, a predicted path segment for a first time interval based on the initial location data, the sensor data, and the location prediction model, the predicted path segment comprising: determining, at the processor, a predicted accuracy of the predicted path segment based on a weight associated with the second sensor and the predicted path segment; determining, at the processor, the predicted accuracy of the predicted path segment exceeds an accuracy threshold; receiving, at the processor from the first sensor, updated location data corresponding to the asset; and transmitting, from a network device of the asset tracking device in communication with the processor to a remote device, the location data, the predicted path segment, the predicted accuracy of the predicted path segment, and the updated location data.

[0026]

[0020] The first sensor can include a GPS sensor and each of the initial location data and the updated location data can include a latitude measurement, a longitude measurement, a heading measurement, a speed measurement, and an accuracy measurement.

[0027]

[0021] The second sensor can include at least one selected from the group of an accelerometer sensor, a temperature sensor, and a gyroscope sensor, and the acceleration data for at least two axis.

[0028]

[0022] A power consumption of a collection of the location data can be higher than a power consumption of a collection of the sensor data.

[0029]

[0023] The predicted path segment can be based on one or more of a predetermined time interval and a distance threshold.

[0030]

[0024] The weight can be determined for the second sensor based on a machine learning model.

[0031]

[0025] The weight can be determined for the second sensor based on at least one selected from the group of: the accuracy measurement of the first sensor, a rate of change for each sensor data, a collection interval of the second sensor.

[0032]

[0026] Transmitting from the network device to the remote device can include transmitting a plurality of predicted path segments, a plurality of initial location data corresponding to each of the predicted path segments, a plurality of predicted accuracies corresponding to each of the plurality of predicted path segments, and a plurality of updated location data corresponding to each of the predicted path segments.

[0033]

[0027] Transmitting from the network device to the remote device can include compressing the plurality of predicted path segments, the plurality of initial location data corresponding to each of the predicted path segments, the plurality of predicted accuracies corresponding to each of the plurality of predicted path segments, and the plurality of updated location data corresponding to each of the predicted path segments.

[0028] In accordance with another aspect, there is provided a computer-implemented method for tracking an asset, comprising: receiving, at a network device from an asset tracking device, a plurality of predicted path segments, a plurality of initial location data corresponding to each of the predicted path segments, a plurality of predicted accuracies corresponding to each of the plurality of predicted path segments, and a plurality of updated location data corresponding to each of the predicted path segments; determining, at a processor in communication with the network device, a plurality of noise-reduced predicted path segments based on the plurality of predicted path segments; and determining, at the processor, based on mapping data, a corresponding street identifier for each of the plurality of noise reduced predicted path segments.

[0034]

[0029] The mapping data can include GIS data.

[0035]

[0030] The method can include: determining, at the processor, for each of the plurality of noise reduced predicted path segments, a revised weight; determining, at the processor, a revised second sensor weight; and transmitting, from the network device to the asset tracking device, the revised second sensor weight.

[0036]

[0031] The weight can be determined for the second sensor based on at least one selected from the group of: the accuracy measurement of the first sensor, a rate of change for each sensor data, a collection interval of the second sensor.

[0037]

[0032] The method can include: determining, at the processor, a dynamic weighting model based on the plurality of noise reduced predicted path segments; and transmitting, from the network device to the asset tracking device, the dynamic weighting model.

[0038]

[0033] It will be appreciated by a person skilled in the art that a system, apparatus, computer-readable medium or method disclosed herein may embody any one or more of the features contained herein and that the features may be used in any particular combination or sub-combination.

[0039]

[0034] These and other aspects and features of various embodiments will be described in greater detail below. Brief Description of the Drawings

[0040]

[0035] For a better understanding of the described embodiments and to show more clearly how they may be carried into effect, reference will now be made, by way of example, to the accompanying drawings in which:

[0041]

[0036]

[0042] FIG. 1 shows a system diagram of a system for tracking an asset using an asset tracking device, in accordance with one or more embodiments.

[0043] FIG. 2 shows a device diagram of an asset tracking device, in accordance with one or more embodiments.

[0044] FIG. 3 shows a method diagram for tracking an asset using an asset tracking device, in accordance with one or more embodiments.

[0045] FIG. 4 shows a method diagram for tracking an asset, in accordance with one or more embodiments.

[0046] FIG. 5 shows an asset tracking diagram, in accordance with one or more embodiments.

[0047]

[0037] The drawings included herewith are for illustrating various examples of articles, methods, and apparatuses of the teaching of the present specification and are not intended to limit the scope of what is taught in any way.

[0048] Description of Exemplary Embodiments

[0049]

[0038] It will be appreciated that numerous specific details are set forth in order to provide a thorough understanding of the example embodiments described herein.

[0050] However, it will be understood by those of ordinary skill in the art that the embodiments described herein may be practiced without these specific details. In other instances, well-known methods, procedures and components have not been described in detail so as not to obscure the embodiments described herein. Furthermore, this description and the drawings are not to be considered as limiting the scope of the embodiments described herein in any way, but rather as merely describing the implementation of the various embodiments described herein.

[0039] It should be noted that terms of degree such as "substantially", "about" and "approximately" when used herein mean a reasonable amount of deviation of the modified term such that the end result is not significantly changed. These terms of degree should be construed as including a deviation of the modified term if this deviation would not negate the meaning of the term it modifies.

[0051]

[0040] In addition, as used herein, the wording “and / or” is intended to represent an inclusive-or. That is, “X and / or Y” is intended to mean X or Y or both, for example. As a further example, “X, Y, and / or Z” is intended to mean X or Y or Z or any combination thereof.

[0052]

[0041] The embodiments of the systems and methods described herein may be implemented in hardware or software, or a combination of both. These embodiments may be implemented in computer programs executing on programmable computers, each computer including at least one processor, a data storage system (including volatile memory or non-volatile memory or other data storage elements or a combination thereof), and at least one communication interface. For example and without limitation, the programmable computers (referred to below as computing devices) may be a server, network appliance, embedded device, computer expansion module, personal computer, laptop, personal data assistant, cellular telephone, smartphone device, tablet computer, wireless device or any other computing device capable of being configured to carry out the methods described herein.

[0053]

[0042] In some embodiments, the communication interface may be a network communication interface. In embodiments in which elements are combined, the communication interface may be a software communication interface, such as those for inter-process communication (IPC). In still other embodiments, there may be a combination of communication interfaces implemented as hardware, software, and a combination thereof.

[0054]

[0043] Program code may be applied to input data to perform the functions described herein and to generate output information. The output information is applied to one or more output devices, in known fashion.

[0055]

[0044] Each program may be implemented in a high level procedural or object-oriented programming and / or scripting language, or both, to communicate with a computer system. However, the programs may be implemented in assembly or machine language, if desired. In any case, the language may be a compiled or interpreted language. Each such computer program may be stored on a storage media or a device (e.g. ROM, magnetic disk, optical disc) readable by a general or special purpose programmable computer, for configuring and operating the computer when the storage media or device is read by the computer to perform the procedures described herein. Embodiments of the system may also be considered to be implemented as a non-transitory computer-readable storage medium, configured with a computer program, where the storage medium so configured causes a computer to operate in a specific and predefined manner to perform the functions described herein.

[0056]

[0045] Furthermore, the systems, processes and methods of the described embodiments are capable of being distributed in a computer program product comprising a computer readable medium that bears computer usable instructions for one or more processors. The medium may be provided in various forms, including one or more diskettes, compact disks, tapes, chips, wireline transmissions, satellite transmissions, internet transmission or downloads, magnetic and electronic storage media, digital and analog signals, and the like. The computer useable instructions may also be in various forms, including compiled and non-compiled code.

[0057]

[0046] Various embodiments have been described herein by way of example only. Various modifications and variations may be made to these example embodiments without departing from the scope of the invention, which is limited only by the appended claims. Also, in the various user interfaces illustrated in the figures, it will be understood that the illustrated user interface text and controls are provided as examples only and are not meant to be limiting. Other suitable user interface elements may be possible.

[0058]

[0047] As recited herein, the term "asset" refers to any form of transportation vehicles including, but not limited to, ships, automobiles, freight trains, motorcycles, buses, trucks, tractor trailers, transportable goods and machinery, and recreational vehicles (RVs). The asset tracking device disclosed herein is designed to operate both as part of an integrated system and as a standalone unit.

[0048] Reference is first made to FIG. 1 , showing a system drawing 100 of a system for tracking an asset using an asset tracking device. The system 100 includes an asset tracking device 102, a network 104, an asset tracking service 106 that includes a server 108 and a database 110, and a user device 112 with a user interface.

[0059]

[0049] The asset tracking device 102 provides for accurately monitoring and reporting the location, movement, and status of various assets. The device 102 is configured to ensure effective asset location tracking and management by providing real-time tracking data and optimizing resource usage.

[0060]

[0050] The asset tracking device 102 may include one or more hardware components. The asset tracking device 102 may include GPS modules for precise location tracking, motion sensors such as accelerometers and gyroscopes for detecting velocity and directional changes, temperature sensors, audio sensors and microcontrollers or processors for data processing. Communication modules, such as cellular or Wi-Fi transceivers, may also be included in the device 102 to enable data transmission to and from the asset tracking service 106. The hardware components may be integrated into the asset tracking device 102, and the device 102 may be mounted on various assets. In an embodiment, the asset tracking device 102 includes a first sensor and a second sensor (not shown). The first sensor can include a GPS sensor that provides initial and updated location data, including latitude, longitude, heading, speed, and accuracy measurements. The second sensor includes at least one sensor from a group consisting of an accelerometer, a temperature sensor, an audio sensor, a pressure sensor, and a gyroscope providing data on acceleration along at least two axes.

[0061]

[0051] The asset tracking device 102 can be configured to acquire sensor data from multiple sources. The device 102 collects location information from GPS satellites, motion data from accelerometers and gyroscopes, and other sensor data such as temperature measurements, audio measurements, pressure measurements, etc. The device 102 can monitor the sensors and acquire data at configurable intervals and / or in response to specific events, such as changes in velocity or direction. The processor (not shown) in the asset tracking device 102 can be configured to receive initial location data from the GPS sensor and sensor data from the accelerometers, gyroscopes, pressure, audio, and temperature sensors. Additionally, the processor can receive updated location data from the GPS sensor at a configurable interval as the asset moves. The data acquisition function provides for maintaining an accurate and up-to-date understanding of the asset’s movements, including in-between the consecutive GPS datapoints.

[0062]

[0052] The power consumption of a collection of the location data may be higher than the power consumption of a collection of the sensor data (as noted, collecting and processing GPS data may be a higher power consumption activity). Therefore, the location data from the GPS sensor, including latitude, longitude, heading, speed, and accuracy measurements, may require more energy compared to collecting data from other sensors like accelerometers, audio sensors, pressure sensors, temperature sensors and gyroscopes. So, to conserve battery life, the processor of the device 102 may be configured to prioritize data collection from the lower-power sensors in a predictable way and estimate the movement of the asset between successive GPS datapoints.

[0063]

[0053] The asset tracking device 102 can be configured to select the sensor data for processing, and to identify when the sensor data is low confidence - requiring an updated collection of location data using GPS. The selection may be guided by machine learning algorithms and predefined rules. Optionally, the processor on the device 102 may first select the active sensor for each behavior using one or more statistical classifiers. The device 102 may select either the first sensor or the second sensor as the active sensor for tracking the asset. For instance, if the processor detects or predicts that the asset is driving on a street, the processor may activate the motion sensors (accelerometer), with the gyroscope, audio sensor, temperature sensor or pressure sensor being optionally activated to provide real-time data for the predictive path between successive GPS locations. In low-speed scenarios, the processor may also enable the e-compass while disabling other sensors such as GPS and LTE connectivity to conserve battery. The selective activation of sensors provides that the asset tracking device 102 uses the relevant data for accurate path prediction while optimizing power usage.

[0054] The asset tracking device 102 can be configured to perform, at the device’s processor, data processing functions to ensure accurate and energy efficient asset tracking. The data collected from various sensors, including the accelerometer, may be processed in real-time by the device 102 using advanced algorithms and predictive models. There may be hundreds or thousands of asset tracking devices 102 reporting asset tracking data to asset tracking service 106.

[0064]

[0055] Predictive algorithms and machine learning refer to computational models and methods to analyze historical and real-time data to make informed predictions about future states, paths, or movements. Predictive algorithms utilize mathematical models to forecast future events based on past data patternsThe predictive algorithms may predict the path, speed, and location of the asset by analyzing data from sensors such as GPS, accelerometers, and gyroscopes. For example, if an asset has been traveling along a highway at a consistent speed, the predictive algorithm may estimate the future position of the asset between consecutive GPS location datapoints.

[0065]

[0056] Machine learning methods described herein may include training models on large datasets to recognize patterns and improve prediction accuracy over time. Machine learning models may be trained to generate sensor data weights, dynamically adjusting the importance of different sensors based on conditions like signal strength or environmental factors. For instance, a machine learning model can be configured to prioritize accelerometer data over GPS data in poor / limited signal areas such as areas with poor satellite signal or where the sky view is blocked by tall buildings, thereby improving the asset tracking accuracy. Alternatively or in addition, machine learning may be used to detect anomalies in the asset’s movement, such as unexpected stops or deviations from the predicted path, which may trigger alerts or corrective actions.

[0066]

[0057] The present disclosure can include various applications of predictive algorithms and machine learning. Optionally, machine learning methods can be applied to enhance the predictive path model by incorporating environmental data such as traffic patterns and road conditions, thereby improving route accuracy. Alternatively or in addition, predictive algorithms can be applied to anticipate maintenance needs based on the asset's usage patterns and sensor data, enabling proactive maintenance and reducing downtime. The technologies enable the system to provide more reliable and energy efficient asset tracking, adapting to changing conditions and continuously improving performance.

[0067]

[0058] In the systems and methods described herein, a weighting method can be applied to dynamically assigning importance to different types of sensor data based on the sensor data accuracy and / or sensor data relevance at any given time. Weight functions can be computed for each sensor input, such as GPS, accelerometers, and gyroscopes, based on sensor importance factors such as signal strength, rates of change, the time interval between measurements and so forth. For instance, when GPS signal is weak or intermittent, higher weights can be assigned to accelerometer and gyroscope data to maintain accurate tracking. The dynamic weights allow the system prioritize the most reliable data, ensuring accurate path prediction while minimizing unnecessary GPS updates, thereby conserving battery life.

[0068]

[0059] Optionally, the asset tracking device processor can be configured to use machine learning models to determine these weights dynamically. For example, the model can be configured to learn from historical data that accelerometer readings are more reliable in urban environments with tall buildings that obstruct GPS signals. Alternatively, the weights can be adjusted based on predefined rules and real-time analysis, such as giving more importance to GPS data when the device detects that the asset is in an open area with strong satellite signals to get quick GPS fix with minimal power consumption. The adaptive weighting mechanism allows the system to maintain high accuracy and efficiency across various operational contexts, reducing both data usage and energy consumption.

[0069]

[0060] The processor of the asset tracking device 102 can be configured to process sensor data. For instance, the collected data from the accelerometer, along with other sensor inputs, can be fed in real-time into data model algorithms to derive the asset’s path without relying solely on continuous GPS updates. The processing can include predictive path models to anticipate future movements and reduce the frequency of GPS updates, thereby conserving battery life. The models may incorporate one or more of historical sensor data, environmental factors, and machine learning techniques to predict the likely path of the asset over time.

[0061] The device 102 can also perform path prediction by predicting an asset’s path in the form of heading, speed, and distance traveled at each measurement point. By analyzing the sensor data, the device 102 can estimate the future position of the asset based on its current and past movements. The predictive path generation provides for maintaining accurate tracking while minimizing the need for frequent GPS data, thus saving battery life.

[0070]

[0062] The accuracy of the predicted path can be determined as a weight function derived from the predefined sensor accuracy and the accuracy of each path attribute calculation, such as heading and speed. The weight function allows the reliability of the predicted path to be assessed. The device 102 can cache the path data such as heading and speed, until further processing is required, providing that the most accurate and reliable path information is used.

[0071]

[0063] The processor of the device 102 can configured to execute a data prediction algorithm that is defined to terminate path prediction upon reaching a predefined confidence threshold for the overall path error or when specific interrupt events occur. For instance, if the predicted path deviates from the actual path or if there is a sudden change in movement, the algorithm can be configured to stop predicting. At this point, the device 102 can capture a full measurement from its sensors, including a high- accuracy GPS fix, motion data, and other relevant telemetry.

[0072]

[0064] When a new location measurement is taken, the device 102 can define an event dataset that includes the new location data, sensor measurements, and the cached data from the previous path predictions (i.e. the intervening route segment). The assembled event dataset can then be transmitted to the recipient server at the asset tracking service 106 for further processing. The device 102 and / or the asset tracking service 106 may perform processing of the event dataset such as filtering and smoothing some data or path prediction attributes, ensuring that only the most relevant and accurate data is used. Optionally, a data optimization process can be used that includes assembling the sensor data into a time series or Fourier function for all the computed attributes. For example, differences in time, heading, and speed can calculated along with the computed accuracy. The improved data can then be used to provide precise and reliable tracking information.

[0065] Optionally, the processor of device 102 can be configured to dynamically adjust the weights assigned to different sensor data based on real-time analysis. Dynamic weight adjustment may ensure that the most relevant and accurate data is prioritized, further enhancing the accuracy of the predictive path model. Machine learning algorithms can be used to continuously improve the predictive model based on incoming data, adapting to changing conditions and improving overall tracking performance.

[0073]

[0066] The device 102 may leverage environmental data and historical movement patterns to refine the predictive path model. By incorporating information such as typical traffic patterns, road conditions, and environmental obstacles, the device can make more informed predictions about the asset’s future movements. This data may be transmitted from asset tracking service 106 to device 102. This may provide the predictive model with input data for use across various scenarios.

[0074]

[0067] Optionally, the device 102 may handle signal measurements from nearby devices. By integrating data from multiple sources, the device 102 may cross-verify information and improve the overall accuracy of the tracking system. The multi-source data integration can be particularly advantageous in environments where GPS signals may be weak or unreliable.

[0075]

[0068] In situations where an asset operates in areas with limited connectivity, the device 102 can store the collected data locally and transmit it in batches once a stable network connection is available with remote service 106. The batch processing can ensure that no data is lost and that the tracking information remains accurate and up- to-date. Furthermore, the processor of device 102 may be configured to execute data compression algorithms to minimize the amount of data transmitted over the network. By compressing the data before transmission, the device 102 can reduce the overall data usage and enhances communication efficiency. This can help maintain reliable tracking in data-constrained environments.

[0076]

[0069] Optionally, the predictive path model executed by the device 102 may be continuously updated and refined based on feedback from the asset tracking service 106. When the asset tracking service 106 processes the transmitted data and detects any deviations or inaccuracies, the server 108 of asset tracking service 106 may send updated predictive models and recalibrated sensor weights back to the device 102. This feedback loop may provide that the device 102 remains accurate and energy efficient in its tracking capabilities.

[0077]

[0070] Optionally, the processor of the device 102 may be configured to determine a predicted path segment for a first-time interval based on initial location data, sensor data, and the location prediction model. The operation includes collecting initial location data from the GPS sensor. The location data includes latitude, longitude, heading, speed, and accuracy measurements. At generally the same time, sensor data from accelerometers and gyroscopes, such as acceleration along multiple axes and angular velocity, may be collected. The location prediction model uses location and sensor date to generate a predicted path segment that estimates the asset’s movement over the first-time interval during the path segment.

[0078]

[0071] Optionally, the processor of the device 102 may be configured to determine a predicted accuracy of the predicted path segment based on a weight associated with the second sensor and the predicted path segment. The processor of the device 102 may compute a weight associated with the second sensor data, which includes, for example, the accelerometer or gyroscope. The weight reflects the reliability and relevance of the sensor data under current conditions. The predicted accuracy is then derived by combining this weight with the predicted path segment data. The predicted path segment data considers factors such as sensor calibration, historical accuracy, and environmental influences. The higher weight indicates more confidence in the predicted path to the system 100.

[0079]

[0072] Optionally, the predicted path segment may be based on one or more of a predetermined time interval and a distance threshold. The device 102 predicts the asset’s path by evaluating the movement over set periods or specific distances traveled. For example, the device 102 may calculate the predicted path every five minutes or after every kilometer traveled, ensuring that the path data is regularly updated without relying on continuous GPS tracking.

[0080]

[0073] Optionally, the weight for the second sensor may be determined based on a machine learning model. The machine learning model continuously learns and adapts from the sensor data, optimizing the weight assigned to the second sensor, based on its reliability and relevance under different conditions. The second sensor may include an accelerometer or gyroscope. The dynamic adjustment improves the accuracy of the predicted path by leveraging the most reliable data sources.

[0081]

[0074] Optionally, the weight for the second sensor may be determined based on at least one selected from the group of: the accuracy measurement of the first sensor, a rate of change for each sensor data, and a collection interval of the second sensor. The weight assigned to the second sensor considers various factors, including the GPS accuracy, the variability in sensor readings, and how frequently the sensor data is collected. The factors considered provide that the most precise and relevant data is used for tracking.

[0082]

[0075] Once the predicted accuracy of the path segment is determined, the processor of the device 102 may be configured to determine if the predicted accuracy of the predicted path segment exceeds an accuracy threshold. The threshold can be defined to ensure that the tracking data meets a certain level of reliability before it is used or transmitted. If the predicted accuracy meets or exceeds the accuracy threshold, the processor executes the next operation. Otherwise, the system 100 may prompt a recalibration or request additional sensor data to improve the prediction accuracy.

[0083]

[0076] to the accuracy threshold can include predefined criteria that the predicted accuracy of a path segment must meet to be considered reliable. The accuracy threshold()s provide that the tracking data is sufficiently accurate for decision-making and transmission to remote systems. For example, a threshold could be set as a specific accuracy percentage, such as 95%, meaning that the predicted path must be within 5% of the actual path to be deemed reliable. Alternatively or in addition, the accuracy threshold(s) may include time-based criteria, where the predicted path must align with real-time data within a certain timeframe, or error margins in distance, where the deviation from the actual path must not exceed a specified number of meters. For instance, in urban environments with dense buildings, the threshold might be stricter due to potential GPS signal interference, requiring the system to rely more on accelerometer and gyroscope data to maintain high accuracy. Conversely, in open rural areas, the threshold might be more lenient due to fewer obstacles affecting GPS signals, allowing for less frequent sensor recalibrations.

[0077] The processor of the device 102 can receive updated location data from the first sensor. The first sensor (e.g. the GPS sensor) can provide location data corresponding to the asset’s new position. This updated data can be used to validate and correct the predicted path segment. By comparing the predicted path with the actual GPS data, the system can adjust the prediction model to enhance future accuracy. This continuous feedback loop can be used to ensure that the predictive model remains accurate and reliable over time.

[0084]

[0078] Optionally, a network module of the asset tracking device 102 transmits the relevant data to a remote device of an asset tracking service. The transmission may include the initial location data, the predicted path segment, the predicted accuracy of the path segment, and the updated location data from the GPS sensor. The data can be sent to a remote server or cloud-based platform, where it is further processed and stored. The remote processing can involve additional analysis, such as integrating GIS information to refine the path accuracy or combining data from multiple devices for comprehensive tracking.

[0085]

[0079] Optionally, the transmitting from the network device of asset tracking device 102 to the remote device of the asset tracking service can include transmitting a plurality of predicted path segments, a plurality of initial location data corresponding to each of the predicted path segments, a plurality of predicted accuracies corresponding to each of the predicted path segments, and a plurality of updated location data corresponding to each of the predicted path segments. The batch transmission operation is executed for comprehensive data transfer, providing the remote device of the asset tracking service with all necessary information to accurately track the asset’s movements.

[0086]

[0080] Optionally, the transmitting from the network device of asset tracking device 102 to the remote device of the asset tracking service can include compressing the plurality of predicted path segments, the plurality of initial location data corresponding to each of the predicted path segments, the plurality of predicted accuracies corresponding to each of the predicted path segments, and the plurality of updated location data corresponding to each of the predicted path segments. By compressing the data, the device 102 reduces the amount of data transmitted, conserving bandwidth and ensuring energy efficient communication while maintaining the integrity and accuracy of the tracking information.

[0087]

[0081] The system 100 can be configured to use different prediction models and / or weighting schemes. For instance, a machine learning model may be employed to continuously improve the weighting mechanism based on new data and environmental changes. The model may learn to prioritize different sensors under various conditions, such as giving more weight to accelerometer data in urban areas with poor GPS signal. Alternatively or in addition, additional sensors may be used, such as temperature or humidity sensors, to improve the prediction accuracy by accounting for environmental factors that might affect the asset's movement.

[0088]

[0082] Optionally, the system 100 may implement a batch processing approach where data is accumulated over a longer period before transmission. The batch processing approach reduces the frequency of data transmissions, conserving bandwidth and reducing network congestion. The device 102 may store several predicted path segments and their corresponding accuracy measures, transmitting them together at predefined intervals or when a stable network connection is available.

[0089]

[0083] The data processing functions of the asset tracking device 102 can be configured to improve battery life, minimize data usage, and enhance the accuracy of asset tracking. By leveraging advanced algorithms, machine learning, and real-time data analysis, the device 102 can provide an energy efficient and reliable tracking solution.

[0090]

[0084] The network 104 includes communication infrastructure for transmitting data between the asset tracking device 102, the asset tracking service 106, and other connected entities. The network 104 may handle the exchange of location data, sensor measurements, predictive path segments, and accuracy information from the device 102 to the asset tracking service 106, and as well, the delivery of new models from the asset tracking service 106 to the asset tracking device 102. The network 104 provides that the tracking data collected by the asset tracking device 102 may be transmitted to the asset tracking service 106 and other remote devices for further processing and analysis.

[0085] The network 104 may include various types of communication networks such as cellular networks (3G, 4G, LTE, and 5G), Wi-Fi networks, satellite communication networks, and other wireless communication technologies like LoRaWAN or Sigfox. The networks may be selected based on the requirement of the mode of data transmission, ranging from high-bandwidth, low-latency cellular and Wi-Fi connections suitable for urban environments, to low-power, long-range networks ideal for rural or remote areas. The choice of network depends on the specific requirements of the tracking application, including coverage area, data transmission frequency, and power consumption considerations.

[0091]

[0086] The network 104 facilitates communication between several key entities. The network 104 connects the asset tracking device 102 with the asset tracking service 106, including the server 108 and the database 110. The connection allows the transmission of collected sensor data, predicted path segments, and accuracy information from the device to the server for processing. Additionally, the network 104 supports communication between the server 108 and the user device 112, which includes a user interface for monitoring and management purposes. Variations and alternatives to the network configuration could include hybrid communication models that leverage multiple network types for redundancy and reliability. For instance, a device could use cellular networks for primary communication and switch to satellite networks in areas with poor cellular coverage. This flexibility ensures continuous and reliable data transmission, regardless of the operational environment.

[0092]

[0087] The asset tracking device 102 can be configured to with a server 108 of an asset tracking service 106 using a communication protocol optimized for efficiency and reliability. The protocol may involve data compression techniques to minimize the size of transmitted data packets and prioritization algorithms to ensure timely delivery of tracking information.

[0093]

[0088] The asset tracking service 106, which can include a server 108 and a database 110, can provide processing, analyzing, and storage of the data collected by the asset tracking device 102. The server 108 can be configured to perform advanced computations and analyses that enhance the accuracy and reliability of asset tracking while offloading computational tasks from the device.

[0089] The server 108 can receive new events from the asset tracking device 102. The new events include the predicted path segments, initial location data, updated location data, and associated accuracy metrics. Upon receiving the new event data, the server 108 can initiate server-side processing to refine and verify the predictive path models. The server-side processing can include cleaning the predicted path data to remove any noise and inaccuracies. This may include regenerating the weight functions for each predicted location and the sensor data, taking into account multiple factors such as GPS accuracy, the rate of change for each sensor data, the interval between measurements, and other predictive factors used for the measurement.

[0094]

[0090] The server 108 can leverage Geographic Information System (GIS) data to enhance the accuracy of the predicted paths. By incorporating GIS information, the server 108 can verify the accuracy of each predicted location and accurately projects the route or street on which the asset is moving. If the prediction falls below a certain threshold of overall weight, the server 108 corrects the path to align with the actual streets or lanes. The GIS-based verification and correction improves the tracking accuracy, especially in complex environments with dense infrastructure.

[0095]

[0091] In cases where GIS information is unavailable, such as in large yards, indoors, or in areas not covered by street maps, the server 108 can apply further corrections to the predicted path by adapting other thresholds for the specific environment. The adaptability provides that the tracking system remains accurate and reliable even in challenging environments where conventional GPS and map data might be insufficient.

[0096]

[0092] The server 108 can be configured to manage a continuous feedback loop with the asset tracking device 102. The feedback loop can include communicating updated predictive paths, recalibrated sensor data weights, and any necessary instructions to adjust the data model on the device. By providing real-time feedback, the server 108 provides that the device 102 may dynamically adjust its tracking behavior based on changing environmental conditions and operational requirements. This continuous interaction between the server and the device enhances the overall efficiency and accuracy of the tracking system.

[0093] Moreover, the server 108 can be configured for long-term data storage and analysis. The database 110 stores all historical tracking data, which may be used for various analytical purposes such as trend analysis, performance monitoring, and predictive maintenance. By analyzing the historical data, the server can identify patterns and anomalies that might indicate potential issues or opportunities for optimization. The database 110 may be a Structured Query Language (SQL) such as PostgreSQL or MySQL or a not only SQL (NoSQL) database such as MongoDB.

[0097]

[0094] Optionally, the server 108 can use machine learning models to continuously improve the predictive path algorithms. The models can be trained on the accumulated data from the database 110 and may adapt to new data inputs to refine the tracking accuracy. For example, the server 108 may learn to better handle specific environmental conditions that affect sensor performance, such as heavy urban infrastructure or rural terrains.

[0098]

[0095] Additionally, the server 108 can be configured to manage the overall network efficiency. By processing data on the server side, the system 100 reduces the amount of data that needs to be transmitted from the asset tracking device 102, conserving bandwidth and reducing the strain on the network 104. This may facilitate scalability and allow the asset tracking devices to operate efficiently with minimal power consumption.

[0099]

[0096] Additionally, the server 108 may integrate data from multiple assets tracking devices, providing a comprehensive view of all tracked assets. The integration can enable more sophisticated analysis and management capabilities, such as fleet management, where the movements and statuses of multiple vehicles or equipment pieces can be monitored and improved in real-time.

[0100]

[0097] By cleaning and refining predictive path data, leveraging GIS information, adapting to various environments, and maintaining a feedback loop with the asset tracking device 102, the server 108 can provide highly accurate and reliable tracking information.

[0101]

[0098] Optionally, the server 108 may be part of an asset tracking service 106 which may be, for example, a cloud-based platform such as Amazon® AWS. The cloudbased platform refers to the infrastructure providing the server 108 and database 110 that implements the asset tracking service 106. The platform leverages cloud computing technologies to offer scalable, flexible, and reliable data processing and storage solutions.

[0102]

[0099] The one or more user devices 112 may be any two-way communication device with capabilities to communicate with other devices that allows users to access the asset tracking information on asset tracking service 106. A user device 112 may be a mobile device such as mobile devices running the Google® Android® operating system or Apple® iOS® operating system, or a personal computer including one running Microsoft® Windows® operating system or Apple® Mac® operating system.

[0103]

[0100] A user device 112 may be the personal device of a user, or may be a device provided by an employer. The one or more user devices 112 may be used by an end user to access the software application (not shown) running on server 108 of asset tracking service 106 over network 104. Optionally, the one or more user devices 112 may access a web application hosted at server 106 using a browser for reviewing asset tracking information for one or more asset tracking devices 102. Alternatively, the one or more user devices 112 may download an application (including downloading from an App Store such as the Apple® App Store or the Google® Play Store) for reviewing asset tracking information on asset tracking service 106. The user device 112 may be a desktop computer, mobile device, or laptop computer. The user device 112 may be in communication with server 106, and may allow a user to review asset tracking information such as route summaries and route segment data stored in database 110, including historical asset tracking information.

[0104]

[0101] The software application running on the one or more user devices 112 may communicate with server 108 of asset tracking service 106 using an Application Programming Interface (API) endpoint, and may send and receive voice sample data, user data, mobile device data, and mobile device metadata.

[0105]

[0102] The software application running on the one or more user devices 112 may display one or more user interfaces on a display device of the user device, including, but not limited to, the user interfaces shown in FIG 1

[0106]

[0103] The user interfaces may allow users to configure and manage the asset tracking devices 102. The feature includes setting parameters for data collection intervals, accuracy thresholds, and communication protocols. By providing the configuration options, the user interface enables users to tailor the tracking system to their specific needs and operational requirements. The user interface may also provide alerts and notifications for events such as deviations from the predicted path, low battery levels, or sensor malfunctions, allowing for proactive management and quick response to potential issues.

[0107]

[0104] By integrating multiple components and functionalities, the system 100 provides for optimizing battery consumption, minimizing data usage, and enhancing accuracy in asset tracking systems. The systems and methods described herein can reduce data or battery consumption while ensuring high-quality data. The device 102 may use a machine learning model to predict data. The device 102 relies on the lowest powerconsuming sensor to reduce battery consumption. Alternatively or in addition, a supporting system or a server 108 can analyze the predicted data and assign an accuracy to it. Alternatively or in addition, the supporting system or a server 108 can apply a weight function and another level of a GIS-aware machine learning model to correct the predictive data. A predictive model can be leveraged to determine the behavior or location of a device instead of relying solely on real sensor measurements. The supporting system or a server 108 may use an alternate predictive model to correct the predictions, resulting in more accurate data. This can reduce the volume of data and battery consumption while maintaining high data quality.

[0108]

[0105] Advantageously, the systems and methods described herein may reduce battery consumption. By dynamically adjusting sensor data weights and incorporating predictive path models, the system 100 can minimize the frequency of GPS updates and improves sensor data usage. This reduces battery consumption compared to traditional systems, extending the operational lifespan of devices and reducing maintenance costs.

[0109]

[0106] The device 102 may provide batched data transfer, to reduce overall network 108 usage. The system 100 can use efficient data compression and transmission protocols, reducing the amount of data transmitted between devices and servers. By prioritizing relevant sensor data and leveraging predictive path models, the system 100 minimizes data traffic without sacrificing tracking accuracy, resulting in lower data usage and reduced operational costs.

[0110]

[0107] The present disclosure also enhances tracking accuracy and reliability. Through the use of predictive path models and server-side recalibration as described above, the system 100 improves tracking accuracy and reliability. By analyzing historical data and considering deviations from expected routes, the system 100 improves location accuracy and reduces errors caused by outdated or inaccurate GPS data.

[0111]

[0108] The device 102 may offer cost efficiency. The system improves resource usage and reduces operational costs associated with battery replacement, data transmission, and maintenance. By minimizing data usage and maximizing the efficiency of sensor data processing, the present solution offers a cost-effective approach for asset tracking applications.

[0112]

[0109] The systems and methods described herein can include both onboard processing within the asset tracking device 102 and server-side processing in a centralized system or cloud-based platform. Algorithms and techniques can be used to improve data processing, making it more efficient in terms of computational resources and memory usage. By dynamically adjusting sensor data weights and employing predictive path models, the computational burden on the device’s processor can be reduced, improving overall system performance and efficiency.

[0113]

[0110] Communication protocols designed for efficiency and reliability can be used, reducing the computational overhead associated with data transmission and reception. By compressing data packets, prioritizing necessary information, and minimizing network traffic, communication efficiency can be enhanced and energy consumption reduced.

[0114]

[0111] Memory management techniques can also be applied to handle large volumes of sensor data and predictive path models efficiently, as the use of memory management processes such as data caching, memory allocation optimization, and garbage collection can ensure optimal memory usage and prevent memory-related bottlenecks in the operation of the computer. The systems and methods described herein can also be integrated seamlessly with existing computer systems.

[0112] Reference is made to FIG. 2, showing an asset tracking device 200 according to an embodiment. The device 200 includes a processor 202, a memory 210, and a network device 220. The processor 202 comprises a data acquisition module 204, a sensor data selection module 206, and a sensor data processing module 208. The memory 210 includes sensor data 212, predictive model data 214, processed data storage 216, and configuration and calibration data 218.

[0115]

[0113] The data acquisition module 204 can be configured to collect data from a plurality of sensors integrated within the asset tracking device 200. The sensors include GPS modules, accelerometers, gyroscopes, and potentially other environmental sensors. The data acquisition module 204 can provide for continual and accurate data collection from these sensors. The data acquisition module 204 may receive initial location data from the GPS sensor. The location data includes latitude, longitude, heading, speed, and accuracy measurements. The data acquisition module 204 stores the location data in the sensor data 212 section of the memory 210 for further processing.

[0116]

[0114] The data acquisition module 204 can be configured to also gather motion data from accelerometers and gyroscopes. The motion data includes velocity, acceleration along multiple axes, and angular velocity. This sensor data is also stored in the sensor data 212 section of the memory 210, enabling the processor 202 to access and utilize it for predictive modeling.

[0117]

[0115] Optionally, the data acquisition module 204 can be configured to monitor the sensors continually. The data acquisition module 204 collects data at predefined intervals or in response to specific events, such as changes in the asset’s velocity or direction. The intervals and event triggers may be stored in the configuration and calibration data 218, which guides data acquisition module 204 operation.

[0118]

[0116] The data acquisition module 204 may also handle variations in data collection frequency. For instance, the data acquisition module 204 may increase the frequency of data collection during rapid movement or reduce it during periods of inactivity. The dynamic adjustment is based on predefined rules stored in the configuration and calibration data 218, optimizing the balance between data accuracy and power consumption.

[0117] Additionally, the data acquisition module 204 is configured to synchronize data collection from multiple sensors. The synchronization provides that data from different sources, such as GPS and motion sensors, is collected simultaneously, providing a comprehensive and accurate dataset. The synchronized data is then stored in the sensor data 212 for integrated processing. The data acquisition module 204 may also detect and handle sensor errors or inconsistencies. If a sensor provides data that deviates from expected values, the data acquisition module 204 flags this data for review. Such flagged data is annotated and stored in the sensor data 212, enabling the sensor data processing module 208 to address these inconsistencies.

[0119]

[0118] The data acquisition module 204 is also configured to interact with external data sources. The data acquisition module 204 may receive supplementary data from nearby devices or external systems to enhance the accuracy of the collected sensor data. This additional data is integrated and stored within the sensor data 212 for comprehensive analysis. The data acquisition module 204 may prepare the collected data for processing. The collected data preparation includes formatting and organizing the data into a structure suitable for input into the sensor data selection module 206 and the sensor data processing module 208. The organized data is then stored in the Sensor Data 212, ready for further analysis and predictive modeling.

[0120]

[0119] The data acquisition module 204 may support real-time data streaming. The data acquisition module 204 can be configured to continuously update the sensor data 212 with new sensor inputs, ensuring that the most current data is always available for processing. This real-time capability can ensure accurate and timely asset tracking.

[0121]

[0120] The sensor data selection module 206 can be configured to analyze and prioritize the data collected by the data acquisition module 204. The sensor data selection module 206 may operate on predefined rules and machine learning algorithms to determine which sensor data is most relevant for processing. The sensor data selection module 206 may select the optimal sensor data based on current conditions and operational requirements. For instance, if the asset is moving rapidly, the module 206 may prioritize accelerometer and gyroscope data. This selected data is then stored in the processed data storage 216 for further analysis.

[0121] The sensor data selection module 206 may use machine learning models to dynamically adjust the importance of different sensor data. The models are stored in the predictive model data 214. The models may be continuously updated based on new data inputs. The module 206 applies these models to prioritize the most reliable and accurate sensor data.

[0122]

[0122] The sensor data selection module 206 can be configured to handle multiple sensors simultaneously. The module 206 can evaluate data from GPS, accelerometers, and gyroscopes to determine which sensor provides the most accurate information under current conditions. The selected data can then be processed and stored in the processed data storage 216. The sensor data selection module 206 may also assess the accuracy and reliability of the collected data. The module 206 uses predefined accuracy thresholds stored in the configuration and calibration data 218 to evaluate sensor data. If the data meets or exceeds these thresholds, the data is selected for further processing. The sensor data selection module 206 may identify and filter out erroneous or inconsistent sensor data. The module 206 compares incoming data against historical patterns and expected values to detect anomalies. The filtering process provides that only high-quality data is used for predictive modeling and is stored in the processed data storage 216.

[0123]

[0123] The sensor data selection module 206 can be configured to prioritize data from lower-power sensors when appropriate. To conserve battery life, the module may select data from accelerometers and gyroscopes, that consume lower electric power, over GPS data when sufficient. The prioritization strategy may be based on rules stored in the configuration and calibration data 218. The module 206 is also configured to integrate data from multiple sensors to enhance overall accuracy. The module 206 combines data from GPS, accelerometers, and gyroscopes to create a comprehensive dataset. The integrated data is then stored in the processed data storage 216 for use in predictive modeling.

[0124]

[0124] The sensor data selection module 206 can be configured to adapt to changing environmental conditions. The module 206 can be configured to dynamically adjust the weight of different sensor data based on factors such as signal strength and environmental interference. The adjustments may be guided by machine learning models stored in the predictive model data 214.

[0125]

[0125] The module 206 may also ensure continuous data flow to the sensor data processing module 208. The module 206 can be configured to constantly evaluate and select the most relevant sensor data, ensuring that the processing module has access to accurate and up-to-date information. The selected data can maintain the accuracy of the asset tracking system and can be stored in the processed data storage 216.

[0126]

[0126] The sensor data processing module 208 can be configured to process the selected sensor data from the sensor data selection module 206. The module 208 may leverage advanced algorithms and predictive models to analyze the data and generate accurate tracking information. The module 208 may apply predictive algorithms to the selected sensor data. The algorithms, stored in the predictive model data 214, may predict the asset’s future path based on historical and real-time sensor data. The processed data can then be stored in the processed data storage 216.

[0127]

[0127] The sensor data processing module 208 can be configured to calculate weights for different sensor inputs. The module 208 can be configured to apply predefined criteria and machine learning models to assign weights, enhancing the accuracy of the predictive models. The weights can then be applied to the sensor data to generate more reliable tracking information. The module 208 may also correct predicted paths using real-time data. By comparing predicted paths with actual GPS data, the module 208 adjusts its predictions to improve accuracy. The corrected data can be stored in the processed data storage 216 for further analysis and transmission. The module 208 may also handle data inconsistencies and errors. The module 208 can apply filtering techniques to smooth out noisy data and correct any detected anomalies. This can ensure that the processed data is accurate and reliable, and it is stored in the processed data storage 216.

[0128]

[0128] The sensor data processing module 208 can be configured to improve battery consumption by reducing the frequency of GPS updates. The module 208 may rely on lower-power sensors like accelerometers and gyroscopes for most data processing tasks, reserving GPS updates for when they are most needed. The prioritization strategy is based on rules and models stored in the configuration and calibration data 218.

[0129]

[0129] The module 208 may generate detailed reports on the asset’s movements. The reports include information such as predicted paths, actual routes, and accuracy metrics. The reports may be stored in the processed data storage 216 and may be transmitted to the server 108 for further analysis.

[0130]

[0130] The sensor data processing module 208 may also prepare data for transmission to the server 108. The module 208 may format and compresses the processed data, ensuring reliable communication over the network 104. The prepared data includes detailed tracking information and is stored in the processed data storage 216 until transmission.

[0131]

[0131] The network device 220 is in communication with the processor 202. The network device 220 can include wired or wireless connection capabilities. The network device 220 can include a radio that communicates utilizing CDMA, GSM, GPRS or Bluetooth protocol according to standards such as IEEE 802.11a, 802.11 b, 802.11g, or 802.11 n. The network device 220 can be used by the asset tracking device 200 to communicate with other devices or computers including the asset tracking service.

[0132]

[0132] Reference is made to FIG. 3, showing a method diagram for tracking an asset using an asset tracking device.

[0133]

[0133] At 302, the method includes providing, at a memory of the asset tracking device, a location prediction model. In this step, the location prediction model can be stored in the memory of the device. The model can be defined to use historical data and machine learning algorithms to predict the asset’s future location based on sensor inputs. The location prediction model can enhance the accuracy of the tracking system while minimizing the need for frequent GPS updates.

[0134]

[0134] At 304, the method includes receiving, at a processor of the asset tracking device in communication with the memory, initial location data from a first sensor positioned at the asset. The first sensor, generally a GPS module, can provide initial location data, including latitude, longitude, heading, speed, and accuracy measurements. The initial location data can serve as the starting point for the predictive model and provides for establishing the asset’s current position.

[0135] At 306, the method includes receiving, at the processor from a second sensor positioned at the asset, sensor data corresponding to the asset. The second sensor may be an accelerometer, gyroscope, or another type of motion sensor. The sensor data includes information on acceleration, angular velocity, and other motion-related parameters.

[0135]

[0136] The method can also include determining a predicted path segment for a firsttime interval based on the initial location data, the sensor data, and the location prediction model.

[0136]

[0137] At 308, the method includes determining, at the processor, a predicted accuracy of the predicted path segment based on a weight associated with the second sensor and the predicted path. The processor can calculate the predicted accuracy by considering the weight assigned to the data from the second sensor, such as an accelerometer or gyroscope. The weight reflects the reliability and relevance of the sensor data under current conditions. By combining the weight with the predicted path segment data, the processor can determine the predicted accuracy. The process provides that the most reliable sensor data is used to enhance the accuracy of the predicted path.

[0137]

[0138] At 310, the method includes determining, at the processor, if the predicted accuracy of the predicted path segment exceeds an accuracy threshold. The threshold provides that the tracking data meets a certain level of reliability before it is used or transmitted. If the predicted accuracy meets or exceeds this threshold, the processor proceeds to the process 312. If not, the method 300 may prompt a recalibration or request additional sensor data to improve the prediction accuracy.

[0138]

[0139] At 312, the method includes receiving, at the processor from the first sensor, updated location data corresponding to the asset. The first sensor, typically including a GPS module, can provide updated location data, including latitude, longitude, heading, speed, and accuracy measurements. The updated data can be used to validate and correct the predicted path segment. By comparing the predicted path with the actual GPS data, the processor may adjust the prediction model to enhance future accuracy. The continuous feedback loop provides that the predictive model remains accurate and reliable over time.

[0140] At 314, the method includes transmitting, from a network device of the asset tracking device in communication with the processor to a remote device, the location data, the predicted path segment, the predicted accuracy of the predicted path segment, and the updated location data. The network device may facilitate the transmission of all relevant data to a remote server or cloud-based platform. The relevant data includes the initial location data, the predicted path segment, the predicted accuracy of the path segment, and the updated location data from the GPS sensor. The transmission may include data compression to reduce the size of data packet. The transmitted data provides for further processing and analysis by a remote server.

[0139]

[0141] Optionally, the first sensor can include a GPS sensor. The initial location data and the updated location data includes a latitude measurement, a longitude measurement, a heading measurement, a speed measurement, and an accuracy measurement.

[0140]

[0142] Optionally, the second sensor can include at least one sensor selected from the group of an accelerometer sensor, a temperature sensor, and a gyroscope sensor. The acceleration data includes measurements for at least two axes.

[0141]

[0143] The power consumption of a collection of the location data may be higher than the power consumption of a collection of the sensor data. The GPS sensor data collection including latitude, longitude, heading, speed, and accuracy measurements, may consume more energy compared to collecting data from lower-power sensors like accelerometers and gyroscopes. Therefore, prioritizing data collection from lower- power sensors may conserve battery life.

[0142]

[0144] The predicted path segment can be based on one or more of a predetermined time interval and a distance threshold. The asset’s path can be predicted by evaluating its movement over set periods or specific distances traveled. For instance, the method may calculate the predicted path every five minutes or after every kilometer traveled, ensuring regular updates without relying on continuous GPS tracking.

[0143]

[0145] The weight for the second sensor can be determined based on a machine learning model. The model may continuously learn and adapt from the sensor data, optimizing the weight assigned to the second sensor based on its reliability and relevance under different conditions. The dynamic adjustment improves the accuracy of the predicted path by leveraging the most reliable data sources.

[0144]

[0146] The weight for the second sensor can be determined based on at least one selected from the group of: the accuracy measurement of the first sensor, a rate of change for each sensor data, and a collection interval of the second sensor. The weight assigned to the second sensor considers various factors, including the GPS accuracy, the variability in sensor readings, and the frequency of data collection.

[0145]

[0147] The transmission from the network device to the remote device of the asset tracking service can include transmitting a plurality of predicted path segments, a plurality of initial location data corresponding to each of the predicted path segments, a plurality of predicted accuracies corresponding to each of the plurality of predicted path segments, and a plurality of updated location data corresponding to each of the predicted path segments.

[0146]

[0148] The transmission from the network device to the remote device of the asset tracking service can include compressing the plurality of predicted path segments, the plurality of initial location data corresponding to each of the predicted path segments, the plurality of predicted accuracies corresponding to each of the predicted path segments, and the plurality of updated location data corresponding to each of the predicted path segments. By compressing the data, the method reduces the overall data size, conserving bandwidth and ensuring efficient communication while maintaining the integrity and accuracy of the tracking information.

[0147]

[0149] Reference is made to FIG. 4, showing a method diagram for tracking an asset.

[0148]

[0150] At 402, the method includes receiving, at a network device from an asset tracking device, a plurality of predicted path segments, a plurality of initial location data corresponding to each of the predicted path segments, a plurality of predicted accuracies corresponding to each of the predicted path segments, and a plurality of updated location data corresponding to each of the predicted path segments. The network device can be configured to collect comprehensive tracking data from the asset tracking device. The predicted path segments can be generated based on the sensor data and the location prediction model, while the initial location data provides the starting points for these segments. The predicted accuracies indicate the confidence levels in the predicted paths. The updated location data received from the GPS sensor further helps validate and adjust the predictions.

[0149]

[0151] At 404, the method includes determining, at a processor in communication with the network device, a plurality of noise-reduced predicted path segments based on the plurality of predicted path segments. The processor can be configured to analyze the received predicted path segments to remove any noise or inaccuracies. The noise reduction process can be performed by advanced algorithms to filter out errant data points that could be caused by sensor errors, environmental interference, or other anomalies. By refining the predicted path segments, the processor provides that the tracking data is more accurate and reliable.

[0150]

[0152] At 406, the method includes determining, at the processor, based on mapping data, a corresponding street identifier for each of the plurality of noise-reduced predicted path segments. The processor can use mapping data, such as Geographic Information System (GIS) data, to identify the specific streets or routes corresponding to the noise-reduced predicted path segments. The cleaned path data can be matched to known geographic locations on the map, allowing the system to accurately place the asset on a specific street or route.

[0151]

[0153] The mapping data can include GIS data. Geographic Information System (GIS) data provides detailed and accurate geographical information. The GIS data may allow the specific streets or routes corresponding to the noise-reduced predicted path segments to be determined.

[0152]

[0154] The method can include determining, at the processor, for each of the plurality of noise-reduced predicted path segments, a revised weight. The processor can calculate revised weights for each path segment based on the noise-reduced data.

[0153]

[0155] The method can include determining, at the processor, a revised second sensor weight. The processor can evaluate the performance of the second sensor, such as an accelerometer or gyroscope, and adjust its weight accordingly.

[0154]

[0156] The method can include transmitting, from the network device to the asset tracking device, the revised second sensor weight. The network device can send the updated weight information back to the asset tracking device. The feedback loop can ensure that the device prioritizes the most reliable sensor inputs.

[0157] The weight for the second sensor can be determined based on at least one selected from the group of: the accuracy measurement of the first sensor, a rate of change for each sensor data, and a collection interval of the second sensor. The processor can use various factors to determine the weight of the second sensor. These factors can include the accuracy of the GPS data, the variability in sensor readings, and how frequently the data is collected.

[0155]

[0158] The method can include determining, at the processor, a dynamic weighting model based on the plurality of noise-reduced predicted path segments. The processor can define a dynamic model that adjusts sensor weights in real-time based on the refined path data. The model can be defined to continuously learns and adapts to changing conditions, improving the accuracy and reliability of the asset tracking system.

[0156]

[0159] The method can include transmitting, from the network device to the asset tracking device, the dynamic weighting model. The network device can send the dynamic weighting model back to the asset tracking device.

[0157]

[0160] Referring next to FIG. 5, there is shown an example asset tracking diagram 500 . A vehicle operating on a road 502 tracked using an asset tracking device at position 102a can collect location data from a GPS sensor and transmit the collected data to an asset tracking service 106.

[0158]

[0161] The asset tracking device collects sensor data from a second sensor between asset tracking device position 102a and asset tracking device position 102b. Using the systems and methods described herein (e.g. the method of FIG. 3), the asset tracking device can use the sensor information from the second sensor to construct a route segment 504 between first 102a and second 102b asset tracking device positions. At the second asset tracking position 102b, the collected second sensor data, as well as the location data collected at asset tracking position 102 and the location data collected at asset tracking position 102b can be transmitted to the asset tracking service 106 for processing.

[0159] While the above description describes features of example embodiments, it will be appreciated that some features and / or functions of the described embodiments are susceptible to modification without departing from the spirit and principles of operation of the described embodiments. For example, the various characteristics which are described by means of the represented embodiments or examples may be selectively combined with each other. Accordingly, what has been described above is intended to be illustrative of the claimed concept and non-limiting. It will be understood by persons skilled in the art that other variants and modifications may be made without departing from the scope of the invention as defined in the claims appended hereto. The scope of the claims should not be limited by the preferred embodiments and examples but should be given the broadest interpretation consistent with the description as a whole.

Claims

ClaimsWe claim:

1. A computer-implemented method for tracking an asset using an asset tracking device, comprising:- providing, at a memory of the asset tracking device, a location prediction model;- receiving, at a processor of the asset tracking device in communication with the memory from a first sensor positioned at the asset, initial location data corresponding to the asset;- receiving, at the processor from a second sensor positioned at the asset, sensor data corresponding to the asset;- selecting, at the processor, the second sensor as an active sensor for tracking the asset;- determining, at the processor, a predicted path segment for a first time interval based on the initial location data, the sensor data, and the location prediction model, the predicted path segment determination comprising:- determining, at the processor, a predicted accuracy of the predicted path segment based on a weight associated with the second sensor and the predicted path segment;- determining, at the processor, the predicted accuracy of the predicted path segment exceeds an accuracy threshold;- receiving, at the processor from the first sensor, updated location data corresponding to the asset; and- transmitting, from a network device of the asset tracking device in communication with the processor to a remote device, the location data, the predicted path segment, the predicted accuracy of the predicted path segment, and the updated location data.

2. The method of claim 1 wherein the first sensor comprises a GPS sensor and each of the initial location data and the updated location data comprises a latitude measurement, a longitude measurement, a heading measurement, a speed measurement, and an accuracy measurement.

3. The method of claim 2 wherein the second sensor comprises at least one selected from the group of an accelerometer sensor, a temperature sensor, and a gyroscope sensor, and the acceleration data for at least two axes.

4. The method of claim 3 wherein a power consumption of a collection of the location data is higher than a power consumption of a collection of the sensor data.

5. The method of claim 4 wherein the predicted path segment is based on one or more of a predetermined time interval and a distance threshold.

6. The method of claim 5 wherein the weight is determined for the second sensor based on a machine learning model.

7. The method of claim 6, wherein the weight is determined for the second sensor based on at least one selected from the group of: the accuracy measurement of the first sensor, a rate of change for each sensor data, a collection interval of the second sensor.

8. The method of claim 7 wherein the transmitting from the network device to the remote device comprises transmitting a plurality of predicted path segments, a plurality of initial location data corresponding to each of the predicted path segments, a plurality of predicted accuracies corresponding to each of the plurality of predicted path segments, and a plurality of updated location data corresponding to each of the predicted path segments.

9. The method of 8 wherein the transmitting from the network device to the remote device comprises compressing the plurality of predicted path segments, the plurality of initial location data corresponding to each of the predicted path segments, the plurality of predicted accuracies corresponding to each of the plurality of predicted path segments, and the plurality of updated location data corresponding to each of the predicted path segments.

10. A computer-implemented device for tracking an asset, comprising a memory and a processor configured to perform the method of any one of claims 1 to 9.

11. A computer-readable media for tracking an asset comprising a media with instructions thereupon for configuring a processor to perform the method of any one of claims 1 to 9.

12. A computer-implemented method for tracking an asset, comprising:- receiving, at a network device from an asset tracking device, a plurality of predicted path segments, a plurality of initial location data corresponding to each of the predicted path segments, a plurality of predicted accuracies corresponding to each of the plurality of predicted path segments, and a plurality of updated location data corresponding to each of the predicted path segments;- determining, at a processor in communication with the network device, a plurality of noise-reduced predicted path segments based on the plurality of predicted path segments; and- determining, at the processor, based on mapping data, a corresponding street identifier for each of the plurality of noise reduced predicted path segments.

13. The method of claim 12 wherein the mapping data comprises GIS data.

14. The method of claim 13 further comprising:- determining, at the processor, for each of the plurality of noise reduced predicted path segments, a revised weight;- determining, at the processor, a revised second sensor weight; and- transmitting, from the network device to the asset tracking device, the revised second sensor weight.

15. The method of claim 14, wherein the weight is determined for the second sensor based on at least one selected from the group of: the accuracy measurement of the first sensor, a rate of change for each sensor data, a collection interval of the second sensor.

16. The method of claim 15 further comprising:- determining, at the processor, a dynamic weighting model based on the plurality of noise reduced predicted path segments; and- transmitting, from the network device to the asset tracking device, the dynamic weighting model.

17. A computer-implemented system for tracking an asset, comprising a memory and a processor configured to perform the method of any one of claims 12 to 16.

18. A computer-readable media for tracking an asset comprising a media with instructions thereupon for configuring a processor to perform the method of any one of claims 12 to 16.

Images