DESC:METHOD AND SYSTEM FOR CRASH DETECTION OF VEHICLE(S)
CROSS REFERENCE TO RELATED APPLICATIONS
The present application claims priority from Indian Provisional Patent Application No. 202421050771 filed on 02/07/2024, the entirety of which is incorporated herein by a reference.
TECHNICAL FIELD
The present disclosure generally relates to a vehicle safety and monitoring. Particularly, the present disclosure relates to a system for real-time detection of a vehicle crash event. Furthermore, the present disclosure relates to a method for real-time detection of a vehicle crash event.
BACKGROUND
In modern days, various accident alert systems have been developed and implemented in the automotive industry, the majority of such systems are primarily designed for four-wheeled vehicles. These systems often rely on hardware and sensing mechanisms that are tailored to the structural, operational, and dynamic characteristics of cars and other large vehicles. Consequently, when these systems are used in two-wheeled vehicles, such as motorcycles or scooters, their performance is often unsatisfactory due to the distinct physical behavior and instability characteristics of two-wheelers during both normal operation and crash events. Also, the existing crash alert systems are slightly inaccurate due to their data collection techniques. Moreover, the existing accident alert systems for vehicles are generally integrated with advanced vehicle networks and vehicle instrument cluster systems, and utilize high-cost components, thereby making the system economically impractical for application in two-wheeler segments particularly in markets where affordability is a critical factor.
Therefore, there exists a need for improved system for vehicle crash detection that overcomes the one or more problems associated as set forth above.
SUMMARY
An object of the present disclosure is to provide a system for real-time detection of a vehicle crash event.
Another object of the present disclosure is to provide a method for real-time detection of a vehicle crash event.
In accordance with first aspect of the present disclosure, there is provided a system for real-time detection of a vehicle crash event. The system comprises at least one sensor, a processing unit and a communication module. The at least one sensor is configured to sense at least one parameter. The processing unit is configured to collect data samples of the at least one parameter from the at least one sensor in a sample window, preprocess the collected data samples to determine maximum and minimum values of the at least one parameter and calculate a delta value for partitioning the collected data samples into a plurality of buckets, assign the collected data samples to the plurality of buckets based on the calculated delta value and a dynamic range of the data samples, form a combined function from the assigned plurality of bucket, dynamically calculate entropy for the combined function using a virtual sensor based on Shannon entropy to assess the variability and unpredictability of the at least one parameter and compare the calculated entropy values against predefined thresholds to detect anomalies indicative of a crash event. The communication module is configured to transmit a crash detection trigger to a central server to generate alerts upon detection of the crash event.
The present disclosure provides the system for real-time detection of the vehicle crash event. The system as disclosed by present disclosure provides an advance and cost-effective system for real-time detection of vehicle crash events. Beneficially, the system is capable to differentiate between normal riding conditions and anomalies indicative of a crash event with high accuracy. Beneficially, the system refines detection sensitivity and responsiveness. Additionally, the system ensures the timely notification and potential emergency response. Moreover, the system architecture is modular, computationally lightweight, and adaptable to two-wheeler platforms, allows for low-cost implementation without compromising on detection reliability or system scalability.
In accordance with second aspect of the present disclosure, there is provided a method for real-time detection of a vehicle crash event. The method comprises sensing at least one parameter using at least one sensor, collecting data samples of the at least one parameter from the at least one sensor in a sample window, preprocessing the collected data samples to determine maximum and minimum values of the at least one parameter and calculate a delta value for partitioning the collected data samples into a plurality of buckets, assigning the collected data samples to the plurality of buckets based on the calculated delta value and a dynamic range of the data samples, forming a combined function from the assigned plurality of buckets, dynamically calculating entropy for the combined function using a virtual sensor based on Shannon entropy to assess the variability and unpredictability of the at least one parameter, comparing the calculated entropy values against predefined thresholds to detect anomalies indicative of a crash event and transmitting a crash detection trigger to a central server to generate alerts upon detection of a crash event.
Additional aspects, advantages, features and objects of the present disclosure would be made apparent from the drawings and the detailed description of the illustrative embodiments constructed in conjunction with the appended claims that follow.
It will be appreciated that features of the present disclosure are susceptible to being combined in various combinations without departing from the scope of the present disclosure as defined by the appended claims.
BRIEF DESCRIPTION OF DRAWINGS
The summary above, as well as the following detailed description of illustrative embodiments, is better understood when read in conjunction with the appended drawings. For the purpose of illustrating the present disclosure, exemplary constructions of the disclosure are shown in the drawings. However, the present disclosure is not limited to specific methods and instrumentalities disclosed herein. Moreover, those in the art will understand that the drawings are not to scale. Wherever possible, like elements have been indicated by identical numbers.
Embodiments of the present disclosure will now be described, by way of example only, with reference to the following diagrams wherein:
FIG. 1 illustrates a block diagram of a system for real-time detection of a vehicle crash event, in accordance with an aspect of the present disclosure.
FIG. 2 illustrates a flow chart of a method for real-time detection of a vehicle crash event, in accordance with another aspect of the present disclosure.
In the accompanying drawings, an underlined number is employed to represent an item over which the underlined number is positioned or an item to which the underlined number is adjacent. A non-underlined number relates to an item identified by a line linking the non-underlined number to the item. When a number is non-underlined and accompanied by an associated arrow, the non-underlined number is used to identify a general item at which the arrow is pointing.
DETAILED DESCRIPTION
The following detailed description illustrates embodiments of the present disclosure and ways in which they can be implemented. Although some modes of carrying out the present disclosure have been disclosed, those skilled in the art would recognise that other embodiments for carrying out or practising the present disclosure are also possible.
The description set forth below in connection with the appended drawings is intended as a description of certain embodiments of a system for real-time detection of a vehicle crash event and is not intended to represent the only forms that may be developed or utilised. The description sets forth the various structures and/or functions in connection with the illustrated embodiments; however, it is to be understood that the disclosed embodiments are merely exemplary of the disclosure that may be embodied in various and alternative forms. The figures are not necessarily to scale; some features may be exaggerated or minimised to show details of particular components. Therefore, specific structural and functional details disclosed herein are not to be interpreted as limiting, but merely as a representative basis for teaching one skilled in the art to variously employ the present invention.
While the disclosure is susceptible to various modifications and alternative forms, specific embodiment thereof has been shown by way of example in the drawings and will be described in detail below. It should be understood, however, that it is not intended to limit the disclosure to the particular forms disclosed, but on the contrary, the disclosure is to cover all modifications, equivalents, and alternatives falling within the scope of the disclosure.
The terms “comprise”, “comprises”, “comprising”, “include(s)”, or any other variations thereof, are intended to cover a non-exclusive inclusion, such that a setup, system that comprises a list of components or steps does not include only those components or steps but may include other components or steps not expressly listed or inherent to such setup or system. In other words, one or more elements in a system or apparatus preceded by “comprises... a” does not, without more constraints, preclude the existence of other elements or additional elements in the system or apparatus.
In the following detailed description of the embodiments of the disclosure, reference is made to the accompanying drawings and which are shown by way of illustration specific embodiments in which the disclosure may be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the disclosure, and it is to be understood that other embodiments may be utilized and that changes may be made without departing from the scope of the present disclosure. The following description is, therefore, not to be taken in a limiting sense.
The present disclosure will be described herein below with reference to the accompanying drawings. In the following description, well known functions or constructions are not described in detail since they would obscure the description with unnecessary detail.
As used herein, the terms “electric vehicle”, “EV”, and “EVs” are used interchangeably and refer to any vehicle having stored electrical energy, including the vehicle capable of being charged from an external electrical power source. This may include vehicles having batteries which are exclusively charged from an external power source, as well as hybrid-vehicles which may include batteries capable of being at least partially recharged via an external power source. Additionally, it is to be understood that the ‘electric vehicle’ as used herein includes electric two-wheeler, electric three-wheeler, electric four-wheeler, electric pickup trucks, electric trucks and so forth.
As used herein, the term “vehicle crash event” refers to an unintentional and abrupt incident involving a vehicle that results in a sudden change in the physical state or motion of the vehicle, such as a rapid deceleration, impact, collision with an object or surface, rollover, or a fall, which may pose a risk of injury to the rider or damage to the vehicle. The crash event is characterized by measurable anomalies or irregularities in one or more sensed parameters such as acceleration, velocity, orientation, or electrical signals detected over a short time interval, which deviate from normal operational patterns beyond a predefined threshold. In the two-wheeled vehicles, a crash event may include scenarios such as tipping over, skidding impact, or abrupt loss of stability.
As used herein, the terms “at least one sensor” and “sensor” are used interchangeably and refer to one or more sensing devices that are configured to detect, measure, or monitor one or more physical parameters associated with the operation or condition of a vehicle. The sensor(s) may include, but are not limited to an accelerometer, gyroscope, magnetometer, GPS module, speed sensor, current sensor, voltage sensor, temperature sensor, or any combination thereof. The sensors may be implemented individually or as part of an integrated sensor module and may be configured to generate time-series data representative of vehicle dynamics, electrical characteristics, or environmental factors. The sensor(s) may be embedded in the vehicle structure, mounted on the vehicle chassis, or located at any suitable position to facilitate accurate data acquisition for crash detection or related applications.
As used herein, the term “at least one parameter” and “parameter” are used interchangeably and refer to one or more measurable physical, electrical, or kinematic characteristics associated with the motion or behavior of a vehicle. The parameters may include but are not limited to linear acceleration in one or more axes (e.g., x, y, z directions), angular velocity or rotational rate, vehicle speed, battery current or voltage, gyroscopic orientation, and tilt angle. The selection of specific parameters may vary depending on the system configuration, sensor placement, and operational requirements for accurately detecting anomalies or crash events. The processing of the parameters enables the system to assess variations in dynamic behavior and determine whether such variations correspond to crash conditions.
As used herein, the term “processing unit” refers to any component or set of components configured to execute instructions, process data, and control the operation of one or more elements of the system. Optionally, the processing unit includes, but is not limited to, a microprocessor, a micro-controller, a complex instruction set computing (CISC) microprocessor, a reduced instruction set (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, or any other type of processing circuit. Furthermore, the term “processor” may refer to one or more individual processors, processing devices and various elements associated with a processing device that may be shared by other processing devices. Furthermore, the processing unit may comprise ARM Cortex-M series processors, such as the Cortex-M4 or Cortex-M7, or any similar processor designed to handle real-time tasks with high performance and low power consumption. Furthermore, the processing unit may comprise custom and/or proprietary processors.
As used herein, the term “data samples” refers to discrete units of information representing measured values of one or more parameters collected by at least one sensor over a defined period of time. Each data sample corresponds to a specific reading captured at a particular instance within a sampling window, and may include, but is not limited to, values such as acceleration, velocity, orientation, current, voltage, or other measurable physical or electrical characteristics associated with the operation or state of a vehicle. The data samples are used by the processing unit for analysis, including but not limited to, preprocessing, segmentation, function formation, entropy calculation, and anomaly detection.
As used herein, the term “sample window” refers to a predefined or dynamically adjustable time frame or data segment during which a series of data samples are collected from one or more sensors monitoring one or more parameters of the vehicle. The sample window is characterized by a fixed or variable number of consecutive data points acquired at a predetermined sampling rate, and is used as the basis for performing signal processing operations, statistical analyses, or event detection algorithms. The size of the sample window may be configured to correspond to a specific duration (e.g., one second) or to a specified number of data samples, and may be dynamically varied based on system requirements, operational conditions, or event detection sensitivity.
As used herein, the term “combined function” refers to a computational construct generated by aggregating and integrating data values from multiple vehicle parameters over a defined time window. The combined function represents a synthesized data profile that captures the collective dynamic behavior of the monitored parameters such as speed, battery current, and linear acceleration along a specific axis during the operation of the vehicle. The values from these parameters are first processed and mapped into categorized buckets based on their distribution characteristics, and then mathematically or algorithmically combined to form a unified function. The combined function serves as a basis for further statistical or entropy-based analysis to detect variations, patterns, or anomalies indicative of abnormal vehicle conditions, including crash events. The combined function may be expressed in the form of a vector, time-series signal, or multi-dimensional array, and may be computed in real time using a sliding or adaptive data window.
As used herein, the term “entropy” refers to a quantitative measure of uncertainty, variability, or randomness in a set of data. The entropy is calculated based on Shannon entropy principles and represents the degree of unpredictability or disorder in a stream of sensor data collected from a vehicle. The higher entropy values indicate greater variability in the sensed parameter, while lower values indicate more predictable or stable patterns. The entropy is used as a diagnostic metric by a virtual sensor to detect anomalies in vehicle behavior, such as those associated with a crash event, by evaluating changes in the statistical distribution of the collected data over a defined time window.
As used herein, the term “virtual sensor” refers to a software-based analytical module configured to estimate or derive a non-directly measurable parameter or condition of a system based on one or more directly measured physical parameters obtained from physical sensors. The virtual sensor operates by processing the measured data using mathematical models, statistical techniques, or signal processing algorithms to generate a computed output that represents a synthesized parameter, such as variability, anomaly likelihood, or system behavior trends. The virtual sensor is configured to calculate entropy-related values such as Shannon entropy or entropy differentials over a sliding window of time to assess the unpredictability and stochastic behavior of sensed parameters for detecting crash-related anomalies in a two-wheeler vehicle.
As used herein, the term “Shannon entropy” refers to a quantitative measure of the uncertainty or randomness associated with a set of data values, based on the principles of information theory. Specifically, Shannon entropy is a statistical metric used to determine the level of unpredictability or information content present in a probability distribution of discrete events. For a given set of sensor data, the Shannon entropy is calculated using the formula:
H=-?_(i=1)^n¦p i log log_2?pi
Where, pi represents the probability of occurrence of each unique value or event in the data set, and n is the total number of unique values. The higher entropy values indicate more variability and less predictability, whereas lower entropy values signify greater uniformity or determinism in the data. The Shannon entropy is used to evaluate the dynamic behavior of one or more vehicle parameters to identify patterns or anomalies indicative of a crash event.
As used herein, the term “calculated entropy values” refers to the quantitative measurements derived from the statistical analysis of a set of data samples corresponding to at least one sensed parameter of a vehicle. The calculated entropy values are computed using an entropy calculation method, such as Shannon entropy, to assess the degree of randomness, variability, or unpredictability present in the sensed data over a defined time window. The calculated entropy values serve as indicators of dynamic behavioural changes in the sensed parameter(s) and are utilized by the system to detect deviations from normal operating conditions that may signify a crash event. The entropy values are computed by first forming a probability distribution of the data samples typically through data bucketing or histogram and then applying the entropy formula to quantify the information content or disorder in the signal.
As used herein, the term “predefined thresholds” refers to one or more reference values that are established in advance based on empirical data, statistical analysis, simulation results, or heuristic rules, and are used as comparison benchmarks to evaluate or determine the occurrence of specific events or conditions. The predefined thresholds are set for entropy values or entropy differentials to distinguish normal operational behavior of the vehicle from anomalous behavior indicative of a crash event. The predefined thresholds may be fixed or dynamically adjustable and may vary based on factors such as vehicle type, operating environment, sensor calibration, or historical performance data.
As used herein, the term “communication module” refers to a hardware component, software component, or a combination thereof, configured to enable the transmission and/or reception of data between the system and one or more external entities. The communication module may include, but is not limited to, wireless communication interfaces such as GSM, LTE, 5G, Wi-Fi, Bluetooth, or any other suitable communication protocol that facilitates real-time or near-real-time data exchange. The communication module is configured to transmit a crash detection trigger, along with relevant metadata, to a remote server, emergency response system, or pre-defined recipient device, thereby facilitating timely alert generation and appropriate action following the detection of a crash event.
As used herein, the term “crash detection trigger” refers to a signal or data packet generated by the system upon identifying one or more conditions that meet or exceed predefined thresholds indicative of a vehicle crash event. The crash detection trigger is produced by the processing unit based on an analysis of sensor data, which may include, but is not limited to, variations in acceleration, velocity, battery current, or other measurable parameters. The trigger represents an actionable output that initiates a subsequent operation, such as activating a communication module to transmit alert information to a remotely located server, emergency contact, or external response system. The crash detection trigger may also include metadata such as timestamp, vehicle status, severity index, and location information, depending on system configuration.
As used herein, the term “central server” refers to a remotely located computing system configured to receive, process, store, and/or redistribute data transmitted by the accident alert system. The central server may comprise one or more physical or virtual computing units connected via a network, such as a cloud-based infrastructure, an internet-connected data centre, or a dedicated backend system. The central server may be configured to receive crash detection triggers, generate accident alerts, log event data, initiate emergency response protocols, notify designated contacts or authorities, and perform analytics or diagnostic evaluations based on the received data. The central server may also interface with other systems, such as mobile devices, emergency services, or vehicular databases, to execute follow-up actions based on the received crash information.
As used herein, the term “alerts” refers to one or more signals, messages, or notifications generated by the system in response to a detected event or condition, such as a crash event. The alerts may be transmitted in various forms, including but not limited to, electronic messages, SMS, push notifications, emails, audible alarms, or visual indicators. These alerts may be directed to one or more recipient entities, such as a remote server, emergency services, a pre-configured contact number, or a mobile application associated with the vehicle or user, to inform them of the occurrence of the detected event and to initiate appropriate follow-up actions.
As used herein, the term “predetermined sampling rate” refers to a fixed or dynamically set frequency at which data samples are collected from one or more sensors over a defined period of time. The sampling rate is established prior to the operation of the system or is adaptively adjusted based on system requirements or detected conditions. It is typically expressed in units such as Hertz (Hz), indicating the number of samples collected per second, and is selected to ensure adequate resolution and accuracy in capturing the variations of the monitored parameter(s) for reliable analysis and event detection.
As used herein, the term “stochasticity” refers to the inherent randomness or variability observed in a sequence of data samples collected from one or more sensors. The stochasticity represents the degree to which the data values deviate unpredictably from a deterministic pattern, and may be quantified by statistical measures such as variance, entropy, or probabilistic distribution characteristics. The system utilizes the concept of stochasticity to dynamically assess fluctuations in sensor data, which aids in distinguishing between normal operating conditions and abnormal or erratic behaviours, such as those associated with a vehicle crash event.
As used herein, the term “data skewness” refers to a statistical characteristic of a dataset in which the distribution of data samples is asymmetrical around the mean. Specifically, the data skewness indicates the extent to which a dataset deviates from a normal (symmetrical) distribution. The data skewness is identified when a significant concentration of data samples occurs within a limited range of values, resulting in an uneven distribution across the partitioned data buckets. The presence of skewness affects the entropy calculation and, therefore, necessitates dynamic adjustment of bucket sizes to improve the accuracy of anomaly detection.
As used herein, the term “entropy differential” refers to a quantitative measure representing the change in entropy values calculated over successive or overlapping time windows of sensor data. Specifically, the entropy differential is determined by computing the average entropy over a predetermined number of consecutive data windows typically three and then evaluating the deviation of this average entropy from one or more predefined baseline or threshold entropy values. The entropy differential serves as an indicator of sudden or abnormal variations in the statistical distribution of the monitored parameter(s), such as acceleration or speed, thereby enabling the system to identify patterns that may correspond to a crash or impact event. The entropy differential enhances the system ability to detect non-linear and unpredictable behaviours in real-time by capturing subtle transitions in system dynamics that may not be evident from raw sensor data alone.
As used herein, the term “severity index” refers to a computed value that quantitatively represents the intensity or seriousness of a detected crash event in a vehicle. The severity index is determined by analyzing the magnitude of deviation in entropy values calculated from the at least one sensed vehicle parameter from predefined threshold values. The severity index may be expressed as a function of the entropy differential, wherein the larger deviations correspond to higher severity levels. The severity index enables the system to classify crash events into different categories, such as minor, moderate, or severe, thereby supporting appropriate response actions, including prioritization of alerts, emergency notifications, or post-crash diagnostics.
As used herein, the term “magnitude of entropy deviation” refers to a quantitative measure representing the extent of variation between a calculated entropy value and one or more predefined entropy threshold values. Specifically, the magnitude of entropy deviation is determined by calculating the absolute difference between the entropy value derived from sensor data such as speed, battery current, or acceleration and the corresponding threshold(s) that define normal operational behavior. A higher magnitude indicates a greater level of unpredictability or disturbance in the sensed parameters, which may be indicative of an abnormal event, such as a crash. The magnitude of entropy deviation serves as a key parameter in assessing the severity of an event and in distinguishing between minor disturbances and significant crash events within the processing unit.
Figure 1, in accordance with an embodiment describes a system 100 for real-time detection of a vehicle crash event. The system 100 comprises at least one sensor, a processing unit 104 and a communication module 106. The at least one sensor 102 is configured to sense at least one parameter. The processing unit 104 is configured to collect data samples of the at least one parameter from the at least one sensor 102 in a sample window, preprocess the collected data samples to determine maximum and minimum values of the at least one parameter and calculate a delta value for partitioning the collected data samples into a plurality of buckets, assign the collected data samples to the plurality of buckets based on the calculated delta value and a dynamic range of the data samples, form a combined function from the assigned plurality of bucket, dynamically calculate entropy for the combined function using a virtual sensor based on Shannon entropy to assess the variability and unpredictability of the at least one parameter and compare the calculated entropy values against predefined thresholds to detect anomalies indicative of a crash event. The communication module 106 is configured to transmit a crash detection trigger to a central server 108 to generate alerts upon detection of the crash event.
In an embodiment, the sample window comprises a predetermined number of samples corresponding to one second of data collected at a predetermined sampling rate. The number of samples corresponds to approximately one second of data collected at a predefined sampling rate. For instance, if the sampling rate is set at 100 Hz, the sample window includes 100 consecutive data points collected over one second. Beneficially, the defined sample window enables the processing unit 104 to analyze real-time variations in the sensed parameter(s) over a consistent and relevant time frame, thereby facilitating accurate detection of the anomalies associated with potential crash events. Moreover, the use of a fixed one-second window balances responsiveness and computational efficiency, ensures the system 100 may operate in real time with reliable detection capabilities.
In an embodiment, the processing unit 104 is configured to partition the collected data samples in three buckets. Also, the processing unit 104 is configured to dynamically adjust the size of each bucket based on the stochasticity of the data samples, increasing the size of bucket where data skewness is observed. The partitioning may be based on the range and distribution characteristics of the collected data samples obtained from the at least one sensor 102 monitoring parameters such as acceleration, speed, or battery current. The processing unit 104 dynamically analyzes the stochastic nature of the data such as randomness, irregularity, and variation in value distribution to identify the presence of skewness in the sample set. Upon detecting data skewness in a specific range, the processing unit 104 may be further configured to adjust the size of the corresponding bucket by expanding the range to better capture the variation within that segment. Beneficially, the adaptive bucket sizing enables more accurate representation and grouping of the collected data, thereby improves the entropy calculation and enhances the ability of the system 100 to detect anomalies associated with crash events. Also, the dynamic adjustment of bucket size ensures that the system 100 remains sensitive to varying signal behaviors under different vehicle operating conditions.
In an embodiment, the processing unit 104 is configured to adjust the sample window size dynamically based on system 100 requirements or specific conditions. The sample window refers to the time interval or the number of data samples collected from the at least one sensor 102 for processing and analysis. The processing unit 104 may be modify the size of this sample window in real time depending on various factors, such as the rate of change in the monitored parameters, environmental conditions, vehicle operating state, or resource constraints like available processing power or memory. For instance, during high-speed operation or periods of detected instability, the processing unit 104 may reduce the window size to enable faster response and more granular anomaly detection. Conversely, during stable or low-risk operation, the window size may be increased to reduce computational load and conserve energy. Beneficially, the adaptive sampling mechanism ensures the optimal performance of the crash detection algorithm under varying operational scenarios, thereby enhances both the accuracy and efficiency of the system 100.
In an embodiment, the processing unit 104 is configured to form the combined function by integrating data from at least three vehicle parameters comprising speed, battery current, and acceleration in the x-direction, over a sliding window of time steps. The at least three vehicle parameters may be collected continuously over the sliding window of time steps, wherein the sliding window represents the dynamically advancing segment of time during which the data samples are recorded and processed. The integration of the at least three parameter into the single combined function allows the system 100 to capture multi-dimensional behavioral patterns of the vehicle, thereby enables the more robust and accurate characterization of vehicle dynamics. The sliding window approach ensures that the most recent and relevant data is always used for analysis, thereby improving the responsiveness of the system 100 to sudden changes in driving conditions, such as those occurring during the crash event. Additionally, by fusing the selected parameters into the unified function, the system 100 enhances the ability to detect complex correlations and patterns that are indicative of crash scenarios, while filtering out routine operational variations.
In an embodiment, the processing unit 104 is configured to calculate an entropy differential via the virtual sensor by averaging the entropy values over three consecutive windows and setting thresholds for anomaly detection based on the entropy differential. The virtual sensor may be implemented in the software within the processing unit 104 and is operable to process entropy values calculated from vehicle parameter data. The processing unit 104 calculates the entropy differential by averaging the entropy values over three consecutive sample windows. Each sample window contains data corresponding to a defined period (e.g., one second) and includes time-series values from the at least one sensed vehicle parameter such as acceleration, speed, or battery current. The averaged entropy value serves to smooth out transient fluctuations and better capture sustained changes in data variability. The system 100 then determines the entropy differential by comparing the averaged value to previous averaged values or predefined entropy thresholds. The thresholds may be configured to represent baseline operating conditions of the vehicle under normal and crash-free scenarios. When the calculated entropy differential exceeds the threshold range, the processing unit 104 identifies the difference as an anomaly, which may correspond to the potential crash event. Beneficially, the approach allows the system to dynamically detect subtle yet significant deviations in sensor data behavior, thereby improves the both sensitivity and reliability of the crash event detection in two-wheeled vehicles.
In an embodiment, the processing unit 104 is configured to determine a severity index of the detected crash event based on a magnitude of entropy deviation from the predefined thresholds. The processing unit 104, upon detecting the anomaly indicative of the crash event using entropy-based analysis, calculates the entropy deviation by determining the difference between the calculated entropy value and the corresponding predefined threshold. The magnitude of the deviation is then used to assess the intensity or severity of the crash. For instance, a higher magnitude of entropy deviation corresponds to a more abrupt or violent change in the monitored parameters (such as acceleration, speed, or current), which may indicate a more severe impact. The severity index is thus a derived metric that allows the system 100 to categorize the crash event into different levels (e.g., minor, moderate, or severe), enabling more informed decision-making regarding alert prioritization, emergency response, or data logging. Beneficially, by the detection of severity index, the system 100 ensures that the crash detection and the impact assessment is to be performed in real time which enhances the effectiveness of the accident alert system.
The present disclosure discloses the system 100 for real-time detection of the vehicle crash event. The system 100 as disclosed by present disclosure is advantageous in terms of a highly adaptive and intelligent system for real-time detection of vehicle crash events, particularly suited for low-cost, resource-constrained environments such as two-wheeled vehicles. Beneficially, by employing the entropy-based anomaly detection using the virtual sensor derived from Shannon entropy, the system 100 is able to accurately assess the unpredictability and abrupt variability in sensor data, which are strong indicators of crash events. The accurate assessment of the unpredictability and abrupt variability in the sensor data eliminates the dependency on traditional threshold-based logic, which may not reliably capture complex, nonlinear crash dynamics in lightweight vehicles. Furthermore, the system 100 further improves detection accuracy by partitioning sensor data into dynamically sized buckets, accounting for data skewness and ensures the robust feature representation. Additionally, by integrating multi-parameter data such as speed, battery current, and directional acceleration, the system 100 enhances the contextual understanding and sensitivity to the broader range of crash scenarios. Moreover, the use of entropy differentials over sliding windows allows for early and reliable detection of anomalies, while the dynamic sampling and window adjustment ensures real-time adaptability under varying operational conditions. Furthermore, the system 100 supports the calculation of a severity index, enables the prioritization in emergency response. Subsequently, the integration of the communication module 106 to trigger alerts in real time ensures rapid relay of critical information to central servers 108 or emergency contacts, thereby improving the post-crash response time and potentially saving lives.
In an embodiment, the system 100 for real-time detection of the vehicle crash event. The system 100 comprises the at least one sensor, the processing unit 104 and the communication module 106. The at least one sensor 102 is configured to sense the at least one parameter. The processing unit 104 is configured to collect data samples of the at least one parameter from the at least one sensor 102 in the sample window, preprocess the collected data samples to determine maximum and minimum values of the at least one parameter and calculate the delta value for partitioning the collected data samples into the plurality of buckets, assign the collected data samples to the plurality of buckets based on the calculated delta value and the dynamic range of the data samples, form the combined function from the assigned plurality of bucket, dynamically calculate entropy for the combined function using the virtual sensor based on Shannon entropy to assess the variability and unpredictability of the at least one parameter and compare the calculated entropy values against predefined thresholds to detect anomalies indicative of the crash event. The communication module 106 is configured to transmit the crash detection trigger to the central server 108 to generate alerts upon detection of the crash event. Furthermore, the sample window comprises the predetermined number of samples corresponding to one second of data collected at the predetermined sampling rate. Furthermore, the processing unit 104 is configured to partition the collected data samples in three buckets. Also, the processing unit 104 is configured to dynamically adjust the size of each bucket based on the stochasticity of the data samples, increasing the size of bucket where data skewness is observed. Furthermore, the processing unit 104 is configured to adjust the sample window size dynamically based on system 100 requirements or specific conditions. Furthermore, the processing unit 104 is configured to form the combined function by integrating data from the at least three vehicle parameters comprising speed, battery current, and acceleration in the x-direction, over the sliding window of time steps. Furthermore, the processing unit 104 is configured to calculate the entropy differential via the virtual sensor by averaging the entropy values over three consecutive windows and setting thresholds for anomaly detection based on the entropy differential. Furthermore, the processing unit 104 is configured to determine the severity index of the detected crash event based on the magnitude of entropy deviation from the predefined thresholds.
Figure 2, describes a method 200 for real-time detection of a vehicle crash event. The method 200 starts at step 202 and completes at step 216. At step 202, the method 200 comprises sensing at least one parameter using at least one sensor 102. At step 204, the method 200 comprises collecting data samples of the at least one parameter from the at least one sensor 102 in a sample window. At step 206, the method 200 comprises preprocessing the collected data samples to determine maximum and minimum values of the at least one parameter and calculate a delta value for partitioning the collected data samples into a plurality of buckets. At step 208, the method 200 comprises assigning the collected data samples to the plurality of buckets based on the calculated delta value and a dynamic range of the data samples. At step 210, the method 200 comprises forming a combined function from the assigned plurality of buckets. At step 212, the method 200 comprises dynamically calculating entropy for the combined function using a virtual sensor based on Shannon entropy to assess the variability and unpredictability of the at least one parameter. At step 214, the method 200 comprises comparing the calculated entropy values against predefined thresholds to detect anomalies indicative of a crash event. At step 216, the method 200 comprises transmitting a crash detection trigger to a central server 108 to generate alerts upon detection of a crash event.
In an embodiment, the sample window comprises a predetermined number of samples corresponding to one second of data collected at a predetermined sampling rate.
In an embodiment, partitioning the collected data samples comprises partitioning the collected data samples into three buckets. The method 200 comprises dynamically adjusting the size of each bucket based on the stochasticity of the data samples, increasing the size of a bucket where data skewness is observed.
In an embodiment, the method 200 comprising adjusting the sample window size dynamically based on system 100 requirements or specific conditions.
In an embodiment, forming the combined function comprises integrating data from at least three vehicle parameters, comprising speed, battery current, and acceleration in the x-direction, over a sliding window of time steps.
In an embodiment, dynamically calculating entropy comprises calculating an entropy differential via the virtual sensor by averaging the entropy values over three consecutive windows and setting thresholds for anomaly detection based on the entropy differential.
In an embodiment, the method 200 comprising determining a severity index of the detected crash event based on a magnitude of entropy deviation from the predefined thresholds.
It would be appreciated that all the explanations and embodiments of the portable device 100 also applies mutatis-mutandis to the method 200.
In the description of the present invention, it is also to be noted that, unless otherwise explicitly specified or limited, the terms “disposed,” “mounted,” and “connected” are to be construed broadly, and may for example be fixedly connected, detachably connected, or integrally connected, either mechanically or electrically. They may be connected directly or indirectly through intervening media, or they may be interconnected between two elements. The specific meaning of the above terms in the present invention can be understood in specific cases to those skilled in the art.
Modifications to embodiments and combination of different embodiments of the present disclosure described in the foregoing are possible without departing from the scope of the present disclosure as defined by the accompanying claims. Expressions such as “including”, “comprising”, “incorporating”, “have”, “is” used to describe and claim the present disclosure are intended to be construed in a non- exclusive manner, namely allowing for items, components or elements not explicitly described also to be present. Reference to the singular is also to be construed to relate to the plural where appropriate.
Although embodiments have been described with reference to a number of illustrative embodiments thereof, it should be understood that numerous other modifications and embodiments can be devised by those skilled in the art that will fall within the spirit and scope of the principles of this disclosure. More particularly, various variations and modifications are possible in the component parts and/or arrangements of the subject combination arrangement within the scope of the present disclosure, the drawings and the appended claims. In addition to variations and modifications in the component parts and/or arrangements, alternative uses will also be apparent to those skilled in the art.
,CLAIMS:WE CLAIM:
1. A system (100) for real-time detection of a vehicle crash event, wherein the system (100) comprises:
- at least one sensor (102) configured to sense at least one parameter;
- a processing unit (104) configured to:
- collect data samples of the at least one parameter from the at least one sensor (102) in a sample window;
- preprocess the collected data samples to determine maximum and minimum values of the at least one parameter and calculate a delta value for partitioning the collected data samples into a plurality of buckets;
- assign the collected data samples to the plurality of buckets based on the calculated delta value and a dynamic range of the data samples;
- form a combined function from the assigned plurality of bucket;
- dynamically calculate entropy for the combined function using a virtual sensor based on Shannon entropy to assess the variability and unpredictability of the at least one parameter; and
- compare the calculated entropy values against predefined thresholds to detect anomalies indicative of a crash event; and
- a communication module (106) configured to transmit a crash detection trigger to a central server (108) to generate alerts upon detection of the crash event.
2. The system (100) as claimed in claim 1, wherein the sample window comprises a predetermined number of samples corresponding to one second of data collected at a predetermined sampling rate.
3. The system (100) as claimed in claim 1, wherein the processing unit (104) is configured to partition the collected data samples in three buckets, wherein the processing unit (104) is configured to dynamically adjust the size of each bucket based on the stochasticity of the data samples, increasing the size of bucket where data skewness is observed.
4. The system (100) as claimed in claim 1, wherein the processing unit (104) is configured to adjust the sample window size dynamically based on system (100) requirements or specific conditions.
5. The system (100) as claimed in claim 1, wherein the processing unit (104) is configured to form the combined function by integrating data from at least three vehicle parameters comprising speed, battery current, and acceleration in the x-direction, over a sliding window of time steps.
6. The system (100) as claimed in claim 1, wherein the processing unit (104) is configured to calculate an entropy differential via the virtual sensor by averaging the entropy values over three consecutive windows and setting thresholds for anomaly detection based on the entropy differential.
7. The system as claimed in claim 1, wherein the processing unit (104) is configured to determine a severity index of the detected crash event based on a magnitude of entropy deviation from the predefined thresholds.
8. A method (200) for real-time detection of a vehicle crash event, wherein the method (200) comprises:
- sensing at least one parameter using at least one sensor (102);
- collecting data samples of the at least one parameter from the at least one sensor (102) in a sample window;
- preprocessing the collected data samples to determine maximum and minimum values of the at least one parameter and calculate a delta value for partitioning the collected data samples into a plurality of buckets;
- assigning the collected data samples to the plurality of buckets based on the calculated delta value and a dynamic range of the data samples;
- forming a combined function from the assigned plurality of buckets;
- dynamically calculating entropy for the combined function using a virtual sensor based on Shannon entropy to assess the variability and unpredictability of the at least one parameter;
- comparing the calculated entropy values against predefined thresholds to detect anomalies indicative of a crash event; and
- transmitting a crash detection trigger to a central server (108) to generate alerts upon detection of a crash event.
9. The method (200) as claimed in claim 8, wherein the sample window comprises a predetermined number of samples corresponding to one second of data collected at a predetermined sampling rate.
10. The method (200) as claimed in claim 8, wherein partitioning the collected data samples comprises partitioning the collected data samples into three buckets, and wherein the method (200) comprises dynamically adjusting the size of each bucket based on the stochasticity of the data samples, increasing the size of a bucket where data skewness is observed.
11. The method (200) as claimed in claim 8, comprising adjusting the sample window size dynamically based on system (100) requirements or specific conditions.
12. The method (200) as claimed in claim 8, wherein forming the combined function comprises integrating data from at least three vehicle parameters, comprising speed, battery current, and acceleration in the x-direction, over a sliding window of time steps.
13. The method (200) as claimed in claim 8, wherein dynamically calculating entropy comprises calculating an entropy differential via the virtual sensor by averaging the entropy values over three consecutive windows and setting thresholds for anomaly detection based on the entropy differential.
14. The method (200) as claimed in claim 8, comprising determining a severity index of the detected crash event based on a magnitude of entropy deviation from the predefined thresholds.