Description:PREDICTION OF DRIVING RANGE OF VEHICLE TECHNICAL FIELD [0001] The present subject matter generally relates to distance to time estimation. More particularly, but not exclusively to a method and system for distance to empty (DTE) estimation. BACKGROUND [0002] Currently, vehicles face a challenge due to the limited range of their batteries. Riders often rely on estimates of the distance to empty (DTE) to plan their trips effectively. However, existing methods for estimating the DTE are often simplistic and fail to account for various factors such as driving conditions, battery characteristics, and user behaviour. [0003] Current methods for estimating the range of the vehicles often rely on simplistic models that fail to capture the complexities of real-world driving conditions. Factors such as speed variations, terrain differences, traffic congestion, and weather conditions significantly impact energy consumption and, consequently, driving range. Existing estimation techniques lack the sophistication to adapt to these diverse conditions accurately. Moreover, batteries used in exhibit diverse characteristics, including variations in state of charge (SOC) among individual battery cells and the overall battery pack. Traditional range estimation methods often overlook these variations, leading to inaccurate predictions of remaining driving distance. Moreover, the health and performance of batteries degrade over time, further complicating range estimation. [0004] Conventionally, telematics systems collect real-time data such as, battery status, speed, and environmental conditions, from the vehicles, to estimate driving range. While telematics-based solutions offer real-time data insights, the aforesaid approach may be unable to predict the DTE. [0005] Typically, empirical formulas derive range estimates based on simplified relationships between driving parameters and energy consumption.
3
5
10
15 20 25
30 Such, empirical formulas may provide rough estimates but often oversimplify the complex interactions between driving conditions, battery characteristics, and user behaviour. Further, the empirical formulas lack accuracy and reliability, particularly in diverse driving environments. [0006] Therefore, there is a need in the art for a method and system for distance to empty (DTE) estimation which addresses at least the aforementioned problems and other problems of known art. [0007] Further limitations and disadvantages of conventional and traditional approaches will become apparent to one of skill in the art, through comparison of described systems with some aspects of the present disclosure, as set forth in the remainder of the present application and with reference to the drawings. SUMMARY OF THE INVENTION [0008] According to embodiments illustrated herein, the present invention provides a method and a system for predicting a driving range of a vehicle. In an embodiment, the method comprises receiving, by a processor, a first dataset associated with a set of parameters of the vehicle for a first time duration. The method comprises training, by the processor, a machine learning model based on the received first dataset. The method comprises receiving, by the processor, a second dataset associated with the set of parameters of the vehicle. In an embodiment, the second dataset comprises real-time information associated with the set of parameters. The method further comprises applying, by the processor, the trained machine learning model on the received second dataset. The method further comprises predicting, by the processor, the driving range based on the application of the trained machine learning model on the second dataset. The method further comprises rendering, by the processor, the predicted driving range on the vehicle. [0009] In an embodiment, the system comprises a plurality of sensors integrated within the vehicle. The plurality of sensors is configured to
4
determine first dataset associated with a set of parameters of the vehicle for a
first time duration and second dataset associated with the set of parameters of the vehicle. The determined first dataset is stored in a memory. The second dataset comprises real-time information associated with the set of parameters. The system further comprises a telematics module configured to receive the 5 stored first dataset and the second dataset. The system further comprises a processor configured to train a machine learning model based on the received first dataset. The processor is further configured to apply the trained machine learning model on the second dataset associated with a set of parameters of the vehicle at a first time instant. The driving range is predicted based on the 10 application of the trained machine learning model. The processor is further configured to render the predicted driving range.
[00010]
It is to be understood that both the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the invention, as claimed. 15
BRIEF DESCRIPTION OF THE DRAWINGS
[00011] The details are described with reference to an embodiment of a gasket for a component assembly along with the accompanying diagrams. The same numbers are used throughout the drawings to reference similar 20 features and components.
[00012] Figure 1 exemplarily illustrates a block diagram of a system for distance to empty (DTE) estimation, in accordance with an embodiment of the present disclosure.
[00013] Figure 2 exemplarily illustrates a block diagram of a system for DTE 25 estimation based on a training of a machine learning model, in accordance with an embodiment of the present disclosure.
[00014] Figure 3 exemplarily illustrates an architecture of the machine learning model, in accordance with an embodiment of the present disclosure.
5
5 10
15 20 25 30 [00015] Figure 4 illustrates a flowchart of a method for DTE estimation, in accordance with an embodiment of the present disclosure. DETAILED DESCRIPTION [00016] Exemplary embodiments are described with reference to the accompanying drawings. Wherever convenient, the same reference numbers are used throughout the drawings to refer to the same or like parts. While examples and features of disclosed principles are described herein, modifications, adaptations, and other implementations are possible without departing from the scope of the disclosed embodiments. It is intended that the following detailed description be considered as exemplary only, with the true scope and spirit being indicated by the following claims. [00017] The terms “an embodiment”, “embodiment”, “embodiments”, “the embodiment”, “the embodiments”, “one or more embodiments”, “some embodiments”, and “one embodiment” mean “one or more (but not all) embodiments of the invention(s)” unless expressly specified otherwise. The terms “including”, “comprising”, “having” and variations thereof mean “including but not limited to”, unless expressly specified otherwise. The terms “a”, “an” and “the” mean “one or more”, unless expressly specified otherwise. [00018] The embodiments of the present invention will now be described in detail with reference to a method for distance to empty (DTE) estimation with the accompanying drawings. However, the present invention is not limited to the present embodiments. The present subject matter is further described with reference to accompanying figures. It should be noted that the description and figures merely illustrate principles of the present subject matter. Various arrangements may be devised that, although not explicitly described or shown herein, encompass the principles of the present subject matter. Moreover, all statements herein reciting principles, aspects, and examples of the present subject matter, as well as specific examples thereof, are intended to encompass equivalents thereof.
6
5
10
15 20 25
30 [00019] A person with ordinary skills in the art will appreciate that the systems, modules, and sub-modules have been illustrated and explained to serve as examples and should not be considered limiting in any manner. It will be further appreciated that the variants of the above disclosed system elements, modules, and other features and functions, or alternatives thereof, may be combined to create other different systems or applications. [00020] The present subject matter is described using the method and system for distance to empty (DTE) estimation, whereas the claimed subject matter can be used in any other type of application employing above-mentioned DTE estimation approaches, with required changes and without deviating from the scope of invention. Further, it is intended that the disclosure and examples given herein be considered as exemplary only. [00021] An objective of the present invention is to provide a method for predicting a driving range of a vehicle. The method comprises receiving, by a processor, a first dataset associated with a set of parameters of the vehicle for a first time duration. The method further comprises training, by the processor, a machine learning model based on the received first dataset. The method further comprises receiving, by the processor, a second dataset associated with the set of parameters of the vehicle. The second dataset comprises real-time information associated with the set of parameters. The method further comprises applying, by the processor, the trained machine learning model on the received second dataset. The method further comprises predicting, by the processor, the driving range based on the application of the trained machine learning model on the second dataset. The method further comprises rendering, by the processor, the predicted driving range on the vehicle. [00022] It may be appreciated that conventional methods for estimating range of two-wheeler vehicles often rely on simplistic models that fail to account for varying driving conditions and individual battery characteristics. Factors such as speed, terrain, traffic, weather, and driving style significantly influence an energy consumption of two-wheeler vehicles, making accurate
7
5 10
15
20 25 30 range prediction a complex task. Moreover, discrepancies between predicted and actual range estimates can result in inconvenience and frustration for riders, impacting the overall user experience. [00023] In order to mitigate the aforesaid issues, disclosed is a method for predicting a driving range of a vehicle. The driving range may be a distance that can be traversed by the vehicle before a battery associated with the vehicle is completely discharged or a fuel tank associated with vehicle is empty. For example, in an electric two-wheeler vehicle, the driving range may give an idea to a rider about the distance that the rider can travel based on a current state-of-charge of the battery associated with the vehicle. The driving range may be also termed as a distance to empty (DTE), an empty range, or a distance to go. In an embodiment, the vehicle is a two-wheeler vehicle, a three-wheeler vehicle or a four-wheeler vehicle. [00024] The method further comprises receiving, by a processor, a first dataset associated with a set of parameters of the vehicle for a first time duration. The first dataset may be a historical dataset associated with the set of parameters of a plurality of vehicles for a plurality of riders. [00025] In an embodiment, the set of parameters is one or more of: a state of charge of each cell in a battery pack, an average state of charge of the battery pack, a drive mode status, a current drawn from the battery pack, a voltage drawn from a battery pack, a speedometer reading, an odometer reading, a temperature sensor reading, a location of the vehicle, information associated with terrain of an area, traffic congestion information, weather information, and user preference information. It may be noted that the state of charge (SOC) of any cell in the battery pack may be a charge remaining in the corresponding cell. The SOC of each cell along with an identification number (ID) may be rendered on a display device associated with the vehicle. The average state of charge (SOC) of the battery pack calculates an average remaining capacity of all the cells in the battery pack. The average SOC value may provide an overall indication of the battery pack's health and remaining range of the vehicle. The drive mode status may indicate a driving mode
8
5 10 15 20 25
30 selected by a driver. For example, the drive mode may be an “eco mode” where energy may be saved. Herein, the one or more features of the vehicle may be disabled to save energy. In another example, the drive mode may be a “power mode” where more energy may be consumed than the “eco mode”. Herein, all features of the vehicle may be enabled to provide a rich rider experience. It may be noted that the drive model may influence power drawn from the battery pack. Therefore, the drive mode may impact the driving range of the vehicle. The current drawn from the battery pack may be a current drawn from an electric motor. The voltage drawn from the battery pack may be a voltage drawn from the electric motor. The current drawn from the battery pack and the voltage drawn from the battery pack may be crucial for calculating energy consumption and predicting the remaining driving range. The speedometer reading may indicate a speed of the vehicle. The odometer reading, a temperature sensor reading, a location of the vehicle, information associated with terrain of an area, traffic congestion information, weather information, and user preference information. The odometer reading may indicate a distance traversed by the vehicle. The temperature sensor reading may indicate a temperature of an area. The location of the vehicle may be a geographical location of the vehicle. The information associated with terrain of the area may indicate a geographical terrain of the area. The traffic congestion information may indicate whether the traffic congestion in the area is low, nominal, or high congestion. The weather information may indicate the weather of the area. For example, the weather information may be “rain”, “windy”, “sunny”, and the like. The user preference information may indicate choices of the user. For example, the user may like to traverse more in “eco mode” than in “power mode”. [00026] In an embodiment, the method may further comprise normalizing one or more values from the received first dataset, filtering one or more outliers from the normalized first dataset, and augmenting the filtered first dataset using at least normalized one or more values. In an example, the first dataset for each of a “10” number of vehicles may be stored. The first dataset associated with a vehicle may include approximately “90,000” data points.
9
5
10 15 20 25
30 Each data point may be normalized to a range of “0” to “1”. Thereafter, data processing techniques may be applied to the normalized first dataset. Herein, the one or more outliers may be removed from the normalized first dataset. Further, missing of one or more features of the normalized first dataset may be imputed or removed. Such data processing techniques may help in reducing noise of the normalized first dataset. [00027] The method may further comprise training, by the processor, the machine learning model based on the received first dataset. In an embodiment, the machine learning model may be a deep multi-layer perceptron. The deep multi-layer perceptron comprises a first layer, a second layer, a third layer of nodes. The second layer and the third layer of nodes may be hidden layers of the deep multi-layer perceptron. The first layer may be an input layer of the deep multi-layer perceptron. The fourth layer may be an output layer of the deep multi-layer perceptron. The received first dataset may be provided as an input to the first layer of the machine learning model. The machine learning model may be trained based on the provided input. Thus, the machine learning model may be trained on a dataset of real-world driving data. During a training phase, the received first dataset may be used to update a weight and bias of the deep multi-layer perceptron through the iterative optimization process. The optimization aimed to minimize the prediction error by adjusting parameters of the machine learning model parameter using techniques such as, gradient descent. In an example, “80” percent of the first dataset may be used for training and “20” percent of the first dataset may be used for testing of the machine learning model. In an embodiment, the machine learning model is trained based on the augmented first dataset. The machine learning model may be trained using a mean absolute error (MAE) loss function. The MAE loss function is a measure of the difference between the predicted driving range and an actual driving range. [00028] The method may further comprise receiving, by the processor, a second dataset associated with the set of parameters of the vehicle. The second dataset comprises real-time information associated with the set of
10
5
10 15
20 25
30 parameters. The set of parameters for the vehicle at a first-time instant may be measured. [00029] The method may further comprise applying, by the processor, the trained machine learning model on the received second dataset. Herein, the received second dataset may be provided as an input to the trained machine learning model. [00030] The method may further comprise predicting, by the processor, the driving range based on the application of the trained machine learning model on the second dataset. Based on the received second dataset, the trained machine learning model may predict the driving range. In an example, the predicted driving range is “25” kilometres. That is, the vehicle may be driven for “25” kilometres till the battery pack are discharged. In an embodiment, the driving range is associated with at least one of a power mode or an eco mode. Herein, the vehicle is the two-wheeler vehicle, or the three-wheeler vehicle. [00031] In an embodiment, the method may further comprise applying an error function on the predicted driving range. The method may further comprise fine-tuning the machine learning model based on the application of the error function. In an example, the error function may be the mean absolute error (MAE) loss function. The predicted driving range and the actual driving range may be provided as inputs to the machine learning model. The MAE loss function may determine the difference between the predicted driving range and the actual driving range. The determined difference may be compared with a threshold value. In case, the determined difference is less than the threshold value, the trained machine learning model may provide optimal results. However, the determined difference is greater than the threshold value, the trained machine learning model may be fine-tuned. [00032] The method further comprises dynamically updating the predicted driving range based on a real-time relationship between the set of parameters of the second dataset. In an example, at a first time instant, the machine learning model may predict a first driving range based on the set of
11
5 10
15 20 25 30 parameters associated with the vehicle for the first time instant. At a second time instant, the machine learning model may predict a second driving range based on the set of parameters associated with the vehicle for the second time instant. The second time instant may be greater than the first time instant. Therefore, the second driving range may be lesser than or equal to the first driving range. In case based on a change in the set of parameters from the first time instant to the second time instance, an intermediary predicted second driving range is greater than the first driving range. The intermediary predicted second driving range may be dynamically updated to predict the second driving range that may be lesser than the first driving range. [00033] The method further comprises rendering, by the processor, the predicted driving range on the vehicle. In an embodiment, the rendering of the predicted driving range is on an audio device associated with the vehicle and/or a display device associated with the vehicle. In an example, the predicted driving range may be rendered on the audio device associated with a helmet of the rider in order to notify the rider about the predicted driving range. In another example, the predicted driving range may be rendered on the display device associated with a smart phone of the rider in order to notify the rider about the predicted driving range. In another example, the predicted driving range may be rendered on the display device positioned at a front region of the vehicle. In an embodiment, the predicted driving range may be rendered in a linearly decreasing manner to reduce stress caused to rider. The real-time relationship being based on rider driving patterns and the second dataset received from one or more sensors. [00034] In another aspect, the system for predicting a driving range of a vehicle is provided. The system comprises a plurality of sensors integrated within the vehicle. The plurality of sensors is configured to determine first dataset associated with a set of parameters of the vehicle for a first time duration and second dataset associated with the set of parameters of the vehicle. The determined first dataset is stored in a memory. The second dataset comprises real-time information associated with the set of parameters. The system may further include a telematics module configured to receive the
12
5
10
15 20
25
stored first dataset and the second dataset. The system may further a processor configured to train a machine learning model based on the received first dataset. The processor may be configured to apply the trained machine learning model on the second dataset associated with a set of parameters of the vehicle at a first time instant. The driving range is predicted based on the application of the trained machine learning model. The processor may be configured to render the predicted driving range. [00035] In an embodiment, the set of parameters is one or more of: a state of charge of each cell in a battery pack, an average state of charge of the battery pack, a drive mode status, a current drawn from a power source, a voltage drawn from a battery pack, a speedometer reading, an odometer reading, a temperature sensor reading, a location of the vehicle, information associated with terrain of an area, traffic congestion information, weather information, and user preference information. [00036] In an embodiment, the processor is further configured to dynamically update the predicted driving range based on a real-time relationship between the set of parameters of the second dataset. [00037] In an embodiment, the predicted driving range is rendered in a linearly decreasing manner to reduce stress caused to rider. The real-time relationship is based on rider driving patterns and the second dataset received from one or more sensors. [00038] In an embodiment, the processor is further configured to normalize one or more values from the received first dataset. The processor is further configured to filter one or more outliers from the normalized first dataset. processor is further configured to augment the filtered first dataset using at least normalized one or more values. The machine learning model is trained based on the augmented first dataset. [00039] In an embodiment, the processor is further configured to render the predicted driving range is an audio device associated with the vehicle and/or a display device associated with the vehicle. 30
13
[00040] Figure 1 exemplarily illustrates a block diagram of a system (100) for distance to empty (DTE) estimation, in accordance with an embodiment of the present disclosure. The system 100 includes a telematics control unit (TCU) 102, a vehicle control unit 104, a master control unit (MCU) 106, a battery management system (BMS) 108, and a cluster 110. The TCU 102 may 5 receive the first dataset from the server. The received first dataset may be transmitted to the vehicle control unit 104 for training of the machine learning model. The vehicle control unit 104 may receive data from the MCU 106, the BMS 108, and the cluster to form the second dataset. The second dataset may be transmitted to the TCU 102. The TCU 102 may transmit the second dataset 10 to the server 112 for storage purposes. It may be noted that the server 112 may be a cloud-based server for storing historical driving data and preferences of an owner of the vehicle. Such data may be used to train and improve an accuracy of the machine learning model over time.
[00041] Figure 2 exemplarily illustrates a block diagram of a system (200) 15 for DTE estimation based on a training of a machine learning model, in accordance with an embodiment of the present disclosure. The system 200 comprises the server 112, an individual SOC 202, an average SOC 204, a drive mode 206, a current and voltage 208, an energy consumed 210, an actual vehicle input data 212, a machine learning model 224, and a computed DTE 20 226.The actual vehicle input data 212 may include a drive mode 214, anindividual SOC 216, an average SOC 218, a current and voltage block 220, an energy consumed 222. The individual SOC 202, the average SOC block 204, the drive mode 206, the current and voltage 208, and the energy consumed 210 may be used in creation of the first dataset. The drive mode 25 214, the individual SOC 216, the average SOC 218, the current and voltage 220, an energy consumed 222 may constitute the second dataset. The actual 30 vehicle input data 212 represents data from various sensors and data sources within the vehicle, such as, a speedometer, an odometer, and an ambient temperature sensor. The machine learning model 224 may be based trained on a historical data, such as, the first dataset to predict the distance to empty (DTE) based on the second dataset. The computed DTE 226 may be the
14
predicted
driving range based on an output of the machine learning model. The predicted driving range is crucial for drivers to plan their trips and avoid running out of battery power.
[00042] Figure 3 exemplarily illustrates an architecture of a machine learning model (300), in accordance with an embodiment of the present disclosure. 5 The machine learning model 300 may include a first layer 302, a second layer 304, a third layer 306, and a fourth layer 308. The first layer 302 may be an input layer. The second layer 304 and the third layer 306 may be hidden layers. The fourth layer 308 may be an output layer. The set of parameters such as, SOC, state of energy (SOE), state-of-health (SOH), odometer 10 reading, the DTE in the eco mode, and the DTE in the power mode may be provided as an input to the first layer 302. The machine learning model 300 may predict the DTE in the eco mode and the DTE in the power mode.
[00043] Figure 4 exemplarily illustrates a method (400) for predicting a driving range of the vehicle, in accordance with an embodiment of the present 15
20 25
30 disclosure. The method 400 includes a block 402, block 404, a block 406, a block 408, a block 410, a block 412, a block 414, a block 416, a block 418, a block 420. The method starts at 402. [00044] At 404, an operation of input data collection may be executed. Herein, the processor may be configured to receive the first dataset associated with the set of parameters of the vehicle for the first time duration. The set of parameters is one or more of a state of charge of each cell in a battery pack, an average state of charge of the battery pack, a drive mode status, a current drawn from the battery pack, a voltage drawn from a battery pack, a speedometer reading, an odometer reading, a temperature sensor reading, a location of the vehicle, information associated with terrain of an area, traffic congestion information, weather information, and user preference information. [00045] At 406, an operation of data preprocessing may be executed. The processor may be configured to normalize one or more values from the received first dataset. The processor may be configured to filter one or more
15
outliers from the normalized first dataset. The processor may be configured
to augment the filtered first dataset using at least normalized one or more values. The machine learning model is trained based on the augmented first dataset.
[00046]
At 408, an operation of deep multiplayer perceptron (MLP) model 5
10
15 20 25
creation may be executed. [00047] At 410, an operation of training the model network may be executed. Herein, the processor may be configured to training, by the processor, the machine learning model based on the received first dataset. [00048] At 412, an operation of integration into a vehicle system may be executed. Herein, the trained machine learning model may be integrated into the vehicle system. [00049] At 414, an operation calculating whether actual range value is equal to an estimated value may be performed. Herein, a third dataset may be provided as an input to the trained machine learning model to estimate a driving range. In case an actual driving range and the estimated driving range is dissimilar, then the flow chart moves to the block 410. In case an actual driving range and the estimated driving range is similar, then the flow chart moves to the block 416. In an embodiment, the error function on may be applied on the actual driving range and the estimated driving range. In case an output of the error function is less than the threshold, then the flowchart 400 may move to the block 410. In case an output of the error function is greater than the threshold, then the flowchart 400 may move to the block 416. The machine learning model may be fine-tuned based on the application of the error function. [00050] At 416, an operation of real-time DTE estimation may be performed. The processor may receive the second dataset associated with the set of parameters of the vehicle. The second dataset comprises real-time information associated with the set of parameters. The processor may apply the trained machine learning model on the received second dataset. The 30
16
5
10
15 20 25
processormay predict the driving range based on the application of the trained machine learning model on the second dataset. The processor may render the predicted driving range on the vehicle. [00051] In an embodiment, the processor may render the predicted driving range is on the audio device associated with the vehicle and/or on the display device associated with the vehicle. In an embodiment, the driving range is associated with at least one of: a power mode or an eco mode. The vehicle is a two-wheeler vehicle, or a three-wheeler vehicle. [00052] 416, an operation of an update and maintenance may be performed. The processor may update parameters of the machine learning model based on the predicted DTE. [00053] In an embodiment, the processor may dynamically update the predicted driving range based on the real-time relationship between the set of parameters of the second dataset. In an embodiment, the predicted driving range is rendered in the linearly decreasing manner to reduce stress caused to rider. The real-time relationship is based on rider driving patterns and the second dataset received from one or more sensors. [00054] In a scenario, the machine learning model is trained using the mean absolute error (MAE) loss function. The machine learning model is trained using an Adam optimizer with a learning rate of 0.001. The proposed method is evaluated on a test dataset of approximately twenty percent of dataset. The MAE of the proposed method is “4.292”, which is significantly lower than the MAE of other state-of-the-art methods and higher value of coefficient of determination (R2) indicate better predictive accuracy and performance of the machine learning model on the testing dataset. Thus, the proposes method is able to outperform other state-of-the-art methods and may be used in a variety of applications such as driver assistance systems, trip planning, battery management systems, and charging infrastructure planning. [00055] In a scenario, existing customer data may be stored on the server via the telematics unit. The machine learning model may be trained based on the 30
17
5 10
15 20 25
30 stored data. The stored data may include each BMS pack indicated SOC, DC Bus Voltage available to MCU, DC Bus Current available to MCU, average SOC, energy consumption, drive mode from MCU indicating if it is in eco or power mode Using each BMS pack indicated SOC, the average SOC may be calculated. Energy consumption is calculated using DC bus voltage available to MCU and DC bus current available to MCU. For predicting DTE in ECO mode, average SOC multiplied by 1.05 may be obtained. For example, for “100” percent SOC, to determine the DTE, 100 multiplied by 1.05 which may be 105 Kilometres range. For predicting DTE in power mode for, average SOC may be multiplied by 0.75, for example, for 100% SOC, 100 may be multiplied by 0.75, that is “75” kilometres range may be the DTE. [00056] In order to mitigate aforesaid problems, the present disclosure provides the method and the system for predicting the driving range of the vehicle. The objectives of the proposed method and system aim to enhance accuracy, enable adaptability to driving conditions, enhance user experience, and in the context of predicting the driving range of the vehicle: accuracy improvement. The accuracy of driving range predictions is enhanced for the vehicle by leveraging advanced machine learning algorithms and real-time data analysis. The discrepancies between predicted and actual driving range estimates are reduced, thereby improving user confidence and satisfaction. The machine learning model is capable of adapting to diverse driving conditions, including speed variations, terrain differences, traffic congestion, weather conditions, and user behaviours. The proposed method incorporates individual battery characteristics and driving preferences into the prediction process to account for variations in energy consumption. [00057] The proposed method and the system may enhance the user experience. The proposed method may provide riders with reliable and personalized estimates of the distance to empty (DTE), enabling them to plan their trips effectively and minimize the risk of unexpected battery depletion. The user satisfaction and confidence are enhanced in electric two-wheelers by delivering accurate and actionable range predictions in real-time.
18
5
10
15
20
25 30 [00058] The proposed method and the system may enable data integration and continual learning. The proposed method integrates telematics data from cloud-based servers to access historical driving data, customer preferences, and environmental factors. The proposed method and the system may continuously update and refine the predictive model based on real-world driving patterns and feedback, ensuring its adaptability and reliability over time. [00059] The proposed method and the system may optimize the parameters using of the machine learning model advanced optimization techniques such as gradient descent and mean absolute error (MAE) loss function. The computational efficiency and scalability of the system is improved to handle large volumes of data and accommodate future enhancements. [00060] The proposed method may enable designing a practical and scalable system architecture capable of integrating with existing electric two-wheeler platforms and telematics infrastructure. The proposed method may ensure ease of deployment and maintenance for manufacturers and service providers, facilitating widespread adoption and utilization of the predictive technology. [00061] The present invention is advantageous as clamping force distribution is optimized. the gasket design that optimally distributes clamping force across the mating surfaces is designed which ensures consistent pressure for a tight seal even in areas with lower initial force. [00062] The present invention is advantageous as the disclosed gasket may adapt to part variations. The disclosed component assembly provides a gasket solution that adapts to variations in part configurations, including shape, material, and thickness, offering versatility for different component assemblies. The disclosed component assembly offers greater design freedom for engineers and designers by introducing a gasket with variable thickness, enabling optimization of component designs both functionally and aesthetically. This helps in minimizing design constraints imposed by traditional gaskets with uniform thickness, allowing for the development of lighter, thinner, and more efficient components.
19
5 10
15 20
25 30
[00063] The present invention improves torque application efficiency. Efficient torque application is facilitated during assembly by ensuring that the thicker portions of the gasket make initial contact, preventing over-tightening and ensuring consistent, reliable sealing. [00064] The proposed system and the method may improve efficiency by leveraging advanced machine learning algorithms, such as multi-layer perceptron (MLP) neural networks. The proposed system and the method achieve higher accuracy in predicting the driving range of the vehicle. By analysing diverse driving parameters and individual battery characteristics, the machine learning model can generate more precise distance to empty (DTE) estimates, reducing the likelihood of unexpected battery depletion. [00065] The proposed system and the method may adapt to a wide range of driving conditions, including variations in speed, terrain, traffic, weather, and user behaviours. [00066] The proposed system and the method may incorporate real-time data from sensors and telematics systems, the machine learning model can dynamically adjust its predictions to reflect changing environmental and driving conditions, enhancing its reliability and relevance. The machine learning model considers individual user preferences, driving styles, and battery characteristics to generate personalized DTE estimates for the riders. By tailoring predictions to each rider's unique profile, the proposed system and the method enhances user satisfaction and confidence in trip planning, facilitating a more enjoyable and stress-free riding experience. [00067] The proposed system and the method integrate telematics data to enable the machine learning model to continuously learn and improve over time, incorporating insights from historical driving data and user feedback. Through iterative updates and refinements, the machine learning model can adapt to evolving driving patterns, battery performance, and environmental factors, ensuring its predictive accuracy and relevance in real-world scenarios. [00068] The proposed system and the method employ advanced optimization techniques, such as gradient descent and mean absolute error (MAE) loss
20
5 10
15 20
25
30 function, to optimize model parameters and improve computational efficiency. The scalable system architecture facilitates seamless integration with existing electric two-wheeler platforms and telematics infrastructure, enabling widespread adoption and deployment across diverse markets. [00069] The proposed system and the method offer the riders valuable decision support for trip planning and route optimization by providing real-time DTE estimates based on current driving conditions and battery status. The riders can make informed decisions about charging stops, route selections, and driving modes, enhancing their overall riding experience and reducing the risk of unplanned battery depletion. [00070] The objectives of the claimed invention collectively aim to address the technical challenges associated with distance range estimation and provide a comprehensive solution that improves distance range estimation methods. [00071] In light of the above-mentioned advantages and the technical advancements provided by the disclosed method and the system for the predicting the driving range of the vehicle, the claimed steps as discussed above are not routine, conventional, or well understood in the art, as the claimed steps enable the following solutions to the existing problems in conventional technologies. Further, the claimed steps clearly bring an improvement in the functioning of the configuration itself as the claimed steps provide a technical solution to a technical problem. [00072] A description of an embodiment with several components in communication with another does not imply that all such components are required, On the contrary, a variety of optional components are described to illustrate the wide variety of possible embodiments of the invention. [00073] Finally, the language used in the specification has been principally selected for readability and instructional purposes, and it may not have been selected to delineate or circumscribe the inventive subject matter and is therefore intended that the scope of the invention be limited not by this detailed description, but rather by any claims that issue on an application
21
5
10 15 based here on. Accordingly, the embodiments of the present invention are intended to be illustrative, but not limiting, of the scope of the invention, which is set forth in the following claims. [00074] While various aspects and embodiments have been disclosed herein, other aspects and embodiments will be apparent to those skilled in the art. The various aspects and embodiments disclosed herein are for purposes of illustration and are not intended to be limiting, with the true scope and spirit being indicated by the following claims. [00075] While the present disclosure has been described with reference to certain embodiments, it will be understood by those skilled in the art that various changes may be made, and equivalents may be substituted without departing from the scope of the present disclosure. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the present disclosure without departing from its scope. Therefore, it is intended that the present disclosure not be limited to the particular embodiment disclosed, but that the present disclosure will include all embodiments falling within the scope of the appended claims.
20
22
Reference Numerals:
100 – System
102-Telematics control unit
104– Vehicle control unit
106– Master control unit 5
108-Battery management system
110– Cluster 202-Individual SOC204-Average SOC206-Drive mode10 208-Current and voltage210-Energy consumed212-Actual vehicle input data214-Drive mode216-Individual SOC15 218-Average SOC220-Current and voltage block222-energy consumed224-Machine learning model226-Computed DTE20 302-First layer304-second layer306-Third layer
23
308-Fourth layer , Claims:We Claim:
1.
A method for predicting a driving range of a vehicle, the method 5 comprising:
receiving, by a processor, a first dataset associated with a set of parameters of the vehicle for a first time duration;
training, by the processor, a machine learning model (224) based on the received first dataset; 10
receiving, by the processor, a second dataset associated with the set of parameters of the vehicle, wherein the second dataset comprises real-time information associated with the set of parameters;
applying, by the processor, the trained machine learning model (224)on the received second dataset;15
predicting, by the processor, the driving range based on theapplication of the trained machine learning model (224) on the second dataset; and
rendering, by the processor, the predicted driving range on the vehicle. 20
2.
The method for predicting a driving range of the vehicle as claimed inclaim 1, wherein the set of parameters is one or more of: a state ofcharge of each cell in a battery pack, an average state of charge of thebattery pack, a drive mode status, a current drawn from the battery25 pack, a voltage drawn from a battery pack, a speedometer reading, anodometer reading, a temperature sensor reading, a location of thevehicle, information associated with terrain of an area, trafficcongestion information, weather information, and user preferenceinformation.30
3.The method for predicting a driving range of the vehicle as claimed inclaim 1, the method further comprising dynamically updating the
25
predicted driving range based on
a real-time relationship between the set of parameters of the second dataset.
4.The method for predicting a driving range of the vehicle as claimed inclaim 3, wherein the predicted driving range being rendered in a5 linearly decreasing manner to reduce stress caused to rider, whereinthe real-time relationship being based on rider driving patterns and thesecond dataset received from one or more sensors.
5.
The method for predicting a driving range of the vehicle as claimed in10 claim 1, the method further comprising:
normalizing one or more values from the received first dataset;
filtering one or more outliers from the normalized first dataset; and
augmenting the filtered first dataset using at least normalized 15 one or more values, wherein
the machine learning model (224) is trained based on the augmented first dataset.
6.
The method for predicting a driving range of the vehicle as claimed in20 claim 1, the method further comprising:
applying an error function on the predicted driving range; and
fine-tuning the machine learning model (224) based on the application of the error function.
25
7.
The method for predicting a driving range of the vehicle as claimed inclaim 1, wherein the machine learning model (224) is a deep multi-layer perceptron, wherein the deep multi-layer perceptron comprisesa first layer (302), a second layer (304), a third layer (306), and afourth layer of nodes (308).30
26
8.
The method for predicting a driving range of the vehicle as claimed inclaim 1, wherein the rendering of the predicted driving range is on anaudio device associated with the vehicle and/or on a display deviceassociated with the vehicle.
5
9.
The method for predicting a driving range of the vehicle as claimed inclaim 1, wherein the driving range is associated with at least one of: apower mode or an eco mode, wherein the vehicle is a two-wheelervehicle, or a three-wheeler vehicle.
10
10.
A system for predicting a driving range of a vehicle, comprising:
a plurality of sensors integrated within the vehicle, wherein the plurality of sensors is configured to determine first dataset associated with a set of parameters of the vehicle for a first time duration and second dataset associated with the set of parameters of the vehicle, 15 wherein the determined first dataset is stored in a memory, wherein the second dataset comprises real-time information associated with the set of parameter;
a telematics module (102) configured to receive the stored first dataset and the second dataset; and 20
a processor configured to:
train a machine learning model (224) based on the received first dataset,
apply the trained machine learning model (224) on the second dataset associated with a set of parameters of the 25 vehicle at a first time instant, wherein the driving range is predicted based on the application of the trained machine learning model (224), and
render the predicted driving range.
30
11.
The system for predicting a driving range of a vehicle as claimed inclaim 10, wherein the set of parameters is one or more of: a state of
27
charge of each cell in a battery pack, an average state of charge of the
battery pack, a drive mode status, a current drawn from a power source, a voltage drawn from a battery pack, a speedometer reading, an odometer reading, a temperature sensor reading, a location of the vehicle, information associated with terrain of an area, traffic 5 congestion information, weather information, and user preference information.
12.The system for predicting a driving range of a vehicle as claimed inclaim 10, wherein the processor is further configured to dynamically10 update the predicted driving range based on a real-time relationshipbetween the set of parameters of the second dataset.
13.The system for predicting a driving range of a vehicle as claimed inclaim 12, wherein the predicted driving range is rendered in a linearly15 decreasing manner to reduce stress caused to rider, wherein the real-time relationship is based on rider driving patterns and the seconddataset received from one or more sensors.
14.
The system for predicting a driving range of a vehicle as claimed in20 claim 10, wherein the processor is further configured to:
normalize one or more values from the received first dataset;
filter one or more outliers from the normalized first dataset; and
augment the filtered first dataset using at least normalized one 25 or more values, wherein
the machine learning model (224) is trained based on the augmented first dataset.
15.
The system for predicting a driving range of the vehicle as claimed in30 claim 10, wherein the processor is further configured to render the
28
predicted driving range is an audio device associated with the vehicle
and/or a display device associated with the vehicle.