Vehicle driver fuel efficiency determination and feedback

Abstract

A system and method for vehicle driver fuel efficiency determination and feedback, are provided. The system may include a device fitted to a vehicle to obtain measurement data during trips of the vehicle. The system may include a server and modelling system for postprocessing of the measurement data. The method may include generating a model that determines contributing factors to a difference between an average fuel efficiency across a fleet of vehicles, and a predicted fuel efficiency for a specific trip of a vehicle. The model may be trained such that the predicted fuel efficiency approximates the actual fuel efficiency. The predicted fuel efficiency and the contributing factors may be output from the modelling system. The system may include a feedback system to provide feedback to a driver or a fleet operator to improve fuel efficiency.

Description

[0001] VEHICLE DRIVER FUEL EFFICIENCY DETERMINATION AND FEEDBACK

[0002] CROSS-REFERENCE TO RELATED APPLICATIONS

[0003] This application claims priority from United Kingdom patent application number GB2418408.7 filed on 16 December 2024, which is incorporated by reference herein.

[0004] FIELD

[0005] This disclosure relates to a system and method for determining a fuel efficiency of a vehicle driver and providing feedback. In particular, the system and method relate to determining driver factors that influence fuel efficiency.

[0006] BACKGROUND

[0007] Large commercial vehicles transport products and goods that are used every day. From transporting products to stores, to construction vehicles, large commercial vehicles may consume large quantities of fuel. Due to potentially large distances that these vehicles travel, the costs of fuel may be a significant cost to the vehicle operator.

[0008] In some cases of a transport company, the company may own tens or hundreds of vehicles. Even a small improvement to the fuel efficiency of each vehicle may result in significant costs savings. It is also important to monitor fuel usage as a measure of determining the environmental impact due to carbon emissions. Therefore, monitoring vehicle fuel efficiency has become an important metric for part of a company’s key performance indicators.

[0009] Traditional fuel efficiency measurements provide only an overall figure. This is not sufficient information for a company to make informed decisions aimed at improving the fuel efficiency of their vehicles.

[0010] There is accordingly scope for improvement.

[0011] The preceding discussion of the background is intended only to facilitate an understanding of the present disclosure. It should be appreciated that the discussion is not an acknowledgment or admission that any of the material referred to was part of the common general knowledge in the art as at the priority date of the application. SUMMARY

[0012] In accordance with an aspect of the disclosure there is provided a computer-implemented method for vehicle driver fuel efficiency feedback, comprising: obtaining measurement data of vehicle data and environmental data during operation of the vehicle over a defined time period; inputting the measurement data into a modelling system configured to: output a predicted fuel efficiency from the measurement data; isolate driver factors relating to the predicted fuel efficiency from environmental factors related to the predicted fuel efficiency, wherein each driver factor relates to a different aspect of a driver’s behaviour; and, quantify a contribution of each driver factor to a difference between an average fuel efficiency and the predicted fuel efficiency, wherein quantifying includes assigning a fuel efficiency contribution to each driver factor; and, outputting an analysis of the driver factors and the assigned fuel efficiency contributions over the defined time period, wherein the average fuel efficiency is an average fuel efficiency value determined using measurement data obtained from previous operations of one or more vehicles over defined time periods.

[0013] The method may include isolating mechanical load factors relating to the predicted fuel efficiency. The method may include isolating mass load factors relating to the predicted fuel efficiency. The method may include quantifying a contribution of the mechanical load factors to the difference between the average fuel efficiency and the predicted fuel efficiency. The method may include quantifying a contribution of the weight load factors to the difference between the average fuel efficiency and the predicted fuel efficiency.

[0014] The method may include obtaining a plurality of vehicle data and environmental data from previous operations of one or more vehicles over defined time periods. The method may include preparing a training dataset from the obtained vehicle data and environmental data. Furthermore, the method may include training the modelling system to output the predicted fuel efficiency to approximate an actual fuel efficiency over the defined time period from vehicle data.

[0015] The method may include determining a target fuel efficiency by subtracting the contributions of each driver factor from the predicted fuel efficiency value.

[0016] The method may include aggregating vehicle and environmental data across multiple time periods to provide insights, wherein the insights include information for any one or more of: drivers, vehicles, and vehicle fleets. The modelling system may be configured to perform a clustering analysis and / or a trend analysis of the actual and predicted fuel efficiency values.

[0017] Outputting the predicted fuel efficiency may include determining a distribution of the fuel efficiency around the predicted fuel efficiency value.

[0018] Quantifying the contribution of each driver factor may include assigning a fuel efficiency contribution to different aspects of a driver’s behaviour.

[0019] The method may include providing a telemetry device associated with the vehicle and configured to obtain the vehicle data and environmental data.

[0020] The telemetry device may obtain the vehicle data from a controller area network (CAN) bus of the vehicle.

[0021] The telemetry device may be configured to link a driver of the vehicle to an operation of the vehicle via a driver identifier, wherein the driver identifier is a physical device configured to uniquely identify and authenticate a driver of one or more drivers.

[0022] The environmental data may include any one or more of: global positioning system (GPS) coordinates, elevation data, road type, and road quality.

[0023] The vehicle data may include any one or more of: actual fuel efficiency, vehicle speed, vehicle acceleration, vehicle total time coasting, a vehicle percentage time coasting, vehicle engine revolutions per minute (RPM), fuel efficiency, torque, kilowatt output, gross combination mass, vehicle make, and vehicle type.

[0024] The vehicle data and the environmental data may include a timestamp associated with each measurement.

[0025] Isolating the driver factors relating to fuel efficiency may include isolating a set of pre-defined driver factors, and wherein the driver factors include any one or more of: RPM management, speed management, acceleration management, driver consistency, or any other operator classification.

[0026] Quantifying the contribution may include quantifying the contribution of each environmental factor to the difference between the average fuel efficiency and the predicted fuel efficiency. The method may include providing a graphical output representation of comparison information, wherein the comparison information includes any one or more of: driver-to-driver comparison, vehicle-to-vehicle comparison, vehicle fleet to vehicle fleet comparison.

[0027] The method may include providing a signal to limit the engine RPM or limiting the vehicle speed in response to specific measurement data.

[0028] The modelling system may be configured to receive, as input, a planned route for a vehicle and output a predicted fuel efficiency for the planned route. The planned route may include an elevation profile of the planned route.

[0029] The modelling system may include any one or more of: an artificial intelligence model, a machine learning model, a deep learning model, and a Shapley decomposition model.

[0030] The modelling system may include a machine learning model or a neural network model, and is configured to receive as input the vehicle data and / or environmental data and output a predicted fuel efficiency value.

[0031] The modelling system may include a Shapley decomposition model, configured to receive any one or more of: a model configured to output a predicted fuel efficiency; vehicle data; and environmental data, and configured to output isolated driver factors relating to the predicted fuel efficiency and to quantify the contribution of each driver factor to the difference between the average fuel efficiency and the predicted fuel efficiency.

[0032] In accordance with a further aspect of the disclosure there is provided a system for vehicle driver fuel efficiency feedback, conducted at a server, comprising: a non-transitory computer-readable storage medium; and one or more processors coupled to the non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium comprises program instructions that, when executed on the one or more processors, cause the system to perform operations comprising: obtaining measurement data of vehicle data and environmental data during operation of the vehicle over a defined time period; inputting the measurement data into a modelling system configured to: output a predicted fuel efficiency from the measurement data; isolate driver factors relating to the predicted fuel efficiency from environmental factors related to the predicted fuel efficiency, wherein each driver factor relates to a different aspect of a driver’s behaviour; and, quantify a contribution of each driver factor to a difference between an average fuel efficiency and the predicted fuel efficiency, wherein quantifying includes assigning a fuel efficiency contribution to each driver factor; and, outputting an analysis of the driver factors and the assigned fuel efficiency contributions over the defined time period, wherein the average fuel efficiency is an average fuel efficiency value determined using measurement data obtained from previous operations of one or more vehicles over defined time periods.

[0033] The modelling system may include a machine learning model, configured to receive vehicle data and / or environmental data, and configured to output a predicted fuel efficiency value, and wherein the modelling system includes a Shapley decomposition model, configured to receive vehicle data and / or environmental data, and configured to output isolated driver factors relating to the predicted fuel efficiency and to quantify the contribution of each driver factor to the difference between the average and predicted fuel efficiency

[0034] In accordance with a further aspect of the disclosure there is provided a computer program product for determining a fuel efficiency of a vehicle, comprising: a non-transitory computer- readable storage medium; and one or more processors coupled to the non-transitory computer- readable storage medium, wherein the non-transitory computer-readable storage medium comprises program instructions that, when executed on the one or more processors, cause the system to perform operations comprising: obtaining measurement data of vehicle data and environmental data during operation of the vehicle over a defined time period; inputting the measurement data into a modelling system configured to: output a predicted fuel efficiency from the measurement data; isolate driver factors relating to the predicted fuel efficiency from environmental factors related to the predicted fuel efficiency, wherein each driver factor relates to a different aspect of a driver’s behaviour; and, quantify a contribution of each driver factor to a difference between an average fuel efficiency and the predicted fuel efficiency, wherein quantifying includes assigning a fuel efficiency contribution to each driver factor; and, outputting an analysis of the driver factors and the assigned fuel efficiency contributions over the defined time period, wherein the average fuel efficiency is an average fuel efficiency value determined using measurement data obtained from previous operations of one or more vehicles over defined time periods.

[0035] Further features provide for the computer-readable medium to be a non-transitory computer- readable medium and for the computer-readable program code to be executable by a processing circuit.

[0036] Embodiments of the technology will now be described, by way of example only, with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] In the drawings:

[0038] Figure 1 is a schematic diagram illustrating an example system for vehicle driver fuel efficiency determination and feedback;

[0039] Figure 2 is a swim-lane flow diagram illustrating an example method for determining a vehicle driver fuel efficiency;

[0040] Figure 3 is a flow diagram illustrating an example method for training a modelling system according to aspects of the disclosure;

[0041] Figure 4A is a block diagram illustrating an example system of components of a server according to aspects of the disclosure;

[0042] Figure 4B is a block diagram illustrating an example system of components of a modelling system according to aspects of the disclosure;

[0043] Figure 5A is an illustration of an example graphical display for providing feedback according to aspects of the disclosure;

[0044] Figure 5B is an illustration of an example graphical display for providing a breakdown of fuel savings per driver according to aspects of the disclosure;

[0045] Figure 50 is an illustration of an example graphical display for providing a trend analysis of fuel savings per driver according to aspects of the disclosure; and,

[0046] Figure 6 illustrates an example of a computing device in which various aspects of the disclosure may be implemented.

[0047] DETAILED DESCRIPTION WITH REFERENCE TO THE DRAWINGS

[0048] A system and method for vehicle driver fuel efficiency determination and feedback, are provided. The fuel efficiency of a vehicle may be a valuable indicator of the performance of the vehicle or driver, or even of the efficiency of the route. Fuel efficiency may be used to determine a cost associated with one or more trips that a vehicle makes, or may be used as a measure of the carbon footprint of said trips. Additionally, determining different factors of vehicle efficiency is an important step to improve fuel efficiency. Separating factors that are within one’s control from factors that are not may allow for these factors to be improved. Such factors may be driver factors or environmental factors. For example, topography may significantly influence the fuel efficiency of a vehicle, but is not a variable that can be controlled or altered by either the driver or the operator for a specific route.

[0049] The vehicle may be a large commercial vehicle for transporting items. Alternatively, the vehicle may be a passenger vehicle, bus, truck, or the like. The vehicle may include a fossil-fuel based power unit. However, the present disclosure is not limited to such a power unit, and may include, but is not limited to, electric vehicles, hybrid, or biofuel-based power units.

[0050] Operators of a vehicle or fleet of vehicles may be able obtain information relating to the operation of the vehicle or vehicles. The operator may be a manager or coordinator of a fleet of vehicles responsible for an operation of the fleet. A telemetry device may be placed onto one or more vehicles. The telemetry device may be usable in collecting data over a defined time period. The defined time period may be a trip performed by the vehicle. A trip may include driving a vehicle from one location to another. The term ‘trip’ and ‘defined time period’ may be used interchangeably. Such data may be that of vehicle data and / or environmental data. The vehicle data and environmental data may be combined into measurement data. The vehicle data may be data of the vehicle itself. The environmental data may be any data not of the vehicle itself. The environmental data may be measured or obtained from sensors situated on the vehicle itself. A driver may be identified as being a driver of the vehicle over a defined time period. If multiple drivers have access to a vehicle, the vehicle may record which driver has made a specific trip.

[0051] By rolling out such telemetry devices across an entire fleet of vehicles, operators may obtain vast quantities of information related to the vehicle, drivers and routes. The telemetry device may correlate data obtained for each trip to a driver and / or a route. The telemetry device may correlate data obtained or collected across multiple defined time periods for a vehicle. Insights may be found by applying appropriate processing means to the data collected across multiple vehicles and multiple trips. Such information may be usable to modify routes. Additionally, the information may be usable to improve a driver fuel efficiency or an environmental fuel efficiency.

[0052] Furthermore, the data obtained may be usable to determine an average fuel efficiency across all trips. The vehicle efficiency may any one of: kilometres per litre (km / L) fuel, litre per kilometre (L / km), litre per hour (L / h), kilometre per kilowatt hour energy (km / kWh), or the like. The disclosure uses a vehicle fuel efficiency of km / L but is not limited to this choice. The average fuel efficiency may be an average across all drivers and / or routes. A postprocessing system may be usable to determine insights from the measurement data. The insights may include information for any one or more of: drivers, vehicles, and vehicle fleets. The insights may be obtained from aggregated vehicle data and environmental data across multiple previous time periods. The postprocessing system may include an averaging step. The averaging step may determine the average fuel efficiency across all vehicles, drivers and / or routes (i.e. across some or all trips). The fuel efficiency may be measured in kilometres travelled per litre fuel (km / L) used. In the case of an electric vehicle, the efficiency may be measured in kilometres travelled per kilowatt hour energy (km / kWh) used. The postprocessing system may determine a predicted fuel efficiency for a specific trip. Once the specific trip is completed, an actual fuel efficiency may be obtained. The actual fuel efficiency may be recorded by the vehicle itself.

[0053] The postprocessing system may include a modelling system. The modelling system may be configured to determine a predicted fuel efficiency from the measurement data. The modelling system may be configured to output the predicted fuel efficiency. With sufficient data collected over time, the modelling system may be configured such that the predicted fuel efficiency should tend towards the actual fuel efficiency. For example, the modelling system may receive information related to a potential trip as an input, such as a planned route, and output a predicted fuel efficiency. After a trip is complete, the actual fuel efficiency may be determined from measurements obtained during the trip. Over time as sufficient data is collected from many different trips, a difference between the predicted fuel efficiency and the actual fuel efficiency may decrease as the prediction performance of the modelling system increases.

[0054] The average fuel efficiency may be determined from data comprising varying conditions. In some cases, a trip may be completely uphill and subsequently, has a low fuel efficiency. In other cases, a trip may be completely downhill with a high fuel efficiency. Averaging these values may hide the fact that the fuel efficiency may have been significantly different from the average in the case of a significant uphill or downhill route. The modelling system may determine a difference between the average fuel efficiency and predicted fuel efficiency. The modelling system may determine a distribution around the average fuel efficiency and / or the predicted fuel efficiency. The modelling system may be configured to isolate driver factors that contributed to the difference between the average fuel efficiency and the predicted fuel efficiency. In an example, if a driver had a trip that was mostly uphill and had a lower fuel efficiency than the average, the modelling system may be able to isolate any driver factors that led to the large deviation. In this example, the deviation may have been that the deviation was only due to the altitude variation and not related to driver factors. The modelling system may be configured to isolate mechanical load factors and / or mass load factors that contribute to the difference between the average fuel efficiency and the predicted fuel efficiency. The mechanical load factors may be factors relating to friction of internal vehicle components, such as gears, bearings for shafts or rotating components, rolling resistance of vehicle tyres, and the like. The mechanical load factors may include any one of more of: vehicle type, vehicle age, vehicle mileage, vehicle service internal, period since previous service, vehicle condition monitoring data, vehicle gearing system, vehicle torque characteristics, vehicle engine attributes, vehicle component friction measurement data, tyre rolling resistance measurement data, and the like. For example, the vehicle condition monitoring may include monitoring various aspects of the vehicle, such as oil volume usage between services, engine torque curve changes, and the like.

[0055] The mass load factors may be factors relating to a load of a mass being transported by the vehicle. The mass load factors may include any one or more of: weight bridge measurement data, predicted load data, vehicle on-board load data, and the like. For example, the predicted load data may be, for a vehicle travelling with an empty load-bed to a mine, the vehicle may be assumed to be at maximum mass capacity when leaving said mine. In some examples, the vehicle on-board load data may include mass data obtained from one or more load cells configured on the vehicle to measure a vehicle mass.

[0056] The examples of mechanical load factors and mass load factors provided are example factors and are not intended to be an exhaustive list of factors.

[0057] The effect of the friction of internal vehicle components and the mass being transported are similar. Both the friction of internal components and mass being transported may result in an additional load being placed on an engine of the vehicle, as both scenarios require more power from the engine to overcome the friction or mass load. For example, an engine of a new vehicle having relatively minimal internal resistance of mechanical components and a relatively heavy load being transported may be required to output the same amount of power (and potentially have a similar fuel efficiency) as an older vehicle with a relatively large amount of internal resistance of mechanical components and a relatively light load being transported.

[0058] Isolating the mechanical load factors and / or mass load factors may improve isolating the driver factor. For example, if a vehicle is driving with a heavy mass, or the vehicle is relatively old with a substantial amount of mechanical friction, the driver may need to drive the vehicle at a high revolutions per minute (RPM). The output of the modelling system may be usable to provide feedback to a driver or operator. A driver feedback system may be usable to improve driver performance and improve fuel efficiency across the fleet. Additionally, isolating external factors relating to the vehicle may be usable in diagnosing vehicle problems that may require additional maintenance.

[0059] The driver factors may include any one or more of: RPM management, speed management, acceleration management, driver consistency, or any other operator classification. For example, the driver consistency may include a combination of the vehicle acceleration and a time period over which a driver had driven the vehicle without accelerating.

[0060] Figure 1 is a schematic diagram illustrating a system 100 for vehicle driver fuel efficiency determination and feedback. In an example, the system may include any one or more of: a server 160, a modelling system 180, a database 140, a telemetry device 120, and a driver device 110. The system may interact with a vehicle 102.

[0061] The vehicle 102 may be a vehicle with a power unit. The vehicle may be a passenger or commercial vehicle. The power unit may be any one of but not limited to: a fossil fuel-based power unit, a hybrid power unit, an electric power unit, or a biofuel-based power unit. The vehicle may be or have been driven by a driver 108.

[0062] The vehicle may include one or more sensors 106. The sensors 106 may be configured to measure aspects of the vehicle. Such sensors 106 may be configured to measure vehicle data. The vehicle data may include any one or more of, but not limited to: a vehicle speed, a vehicle acceleration, a vehicle coasting, a vehicle percentage time coasting, an actual fuel efficiency, a vehicle engine revolutions per minute (RPM), fuel efficiency, torque, kilowatt output, gross combination mass, vehicle make, and vehicle type. The vehicle coasting may be a total time that the vehicle is coasting during the defined time period. Coasting may, for example, occur when the vehicle is propelled by gravity or other forces apart from a power source of the vehicle. The vehicle may include a controller area network (CAN) bus 104. The CAN bus may collect signals from the sensors 106. The CAN bus may be provided on vehicles to access information regarding the vehicles. Additionally, the CAN bus may determine other signals based on measured information, such as the fuel efficiency, the torque, and the like. The CAN bus is often found on heavy operating vehicles. The disclosure is not limited to using a CAN bus. The system and method for vehicle driver fuel efficiency determination and feedback may work with any device on the vehicle that is configured to obtain vehicle data.

[0063] The CAN bus may communicate with a telemetry device 120. The telemetry device 120 may be configured to obtain vehicle data and environmental data. The telemetry device may be connected to the CAN bus via a network cable for transmitting data to the telemetry device. The telemetry device may communicate with the CAN bus and / or sensors wirelessly. The telemetry device may include one or more telemetry device sensors 122. The telemetry device sensors 122 may collect environmental data, including any or more of, but not limited to: a global positioning system (GPS) coordinate, elevation data, road type, road quality. The road type may further include: national, regional, and the like. The road quality may include any one or more of: presence of potholes, percentage quality of road, road condition based on accelerometer data, instantaneous road slope, average road slope, road curvature, road material (tar, gravel), and the like. In an example, the road quality may be determined by cameras on the vehicle. The telemetry device sensors 122 may include accelerometers. The accelerometers may be usable to represent a road quality by accounting for sudden vertical acceleration of the vehicle due to bumps or holes in the road.

[0064] Some variables of the vehicle data may be measured directly by the CAN bus. Other variables may be derived and / or augmented from the directly measured data. Additionally, the telemetry device may collect and / or determine some of the vehicle data. Vehicle data may originate from a customer relationship management software system, for example, vehicle type or vehicle model.

[0065] The telemetry device 120 may include an acquisition module 124. The acquisition module 124 may be configured to obtain signals from the CAN bus 104. The acquisition module 124 may be usable to obtain data from external sensors, such as a GPS or a weigh bridge sensor data being a measurement of a mass of the vehicle. The telemetry device 120 may include a data derivation module 126. The data derivation module may be configured to derive data from measurement data. In an example, the data derivation module may determine a vehicle acceleration by determining a derivative of vehicle speed. The telemetry device 120 may include a driver authorisation module 128. The driver authorisation module 128 may be configured to authorise a driver 108. Authorising the driver may include associating any data obtained over the defined time period to the driver until the driver finishes a trip. The telemetry device may include a driver authentication sensor 129. The driver authentication sensor 129 may scan a physical device of the driver.

[0066] The elevation data may include any one or more of: total altitude measurement, total ascent, total descent, average altitude, total rate of elevation change, average rate of elevation change, or the like. The elevation data may include any data that may be derived from the aforementioned list. The data obtained by the sensors 106,122 may be timestamped data. The telemetry device may be configured to be fitted to the vehicle. The server 160 may be configured to interact with various components of the system 100. The server may include a vehicle driver fuel efficiency feedback system 162. The vehicle driver fuel efficiency feedback system 162 may be configured to transmit feedback to the driver 108. The feedback may include driver behaviour feedback for providing feedback on the drivers specific driving behaviour. Additionally, the vehicle driver fuel efficiency feedback system 162 may include an operator feedback module 169, configured to provide feedback to an operator or coordinator of one or more vehicles. The operator feedback module 169 may be configured to provide feedback to the operator at regular intervals or upon request by the operator. The feedback may be transmitted to a control centre of the operator. The operator feedback module 169 may provide graphical representation feedback of the driver, the vehicle and / or the vehicle fleet.

[0067] All modules within the server may form part of the vehicle driver fuel efficiency feedback system 162. The server 160 may include a communication module 163. The communication module 163 may be configured to communicate with any one or more of: the driver device 110, the telemetry device 120, the database 140, and the modelling system 180. The server 160 may include a model interaction module 164. The model interaction module 164 may be configured to interact, such as by receiving and transmitting data between, with the modelling system 180 and the server 160 and / or database 140.

[0068] The modelling system 180 may be a system configured to generate and use a model. Such a system may be provided by a third party, such as Jupyter Notebooks™, PyTorch™, or Tensorflow™. The modelling system 180 may be hosted by a third-party and accessible via the communication module 163. Alternatively, the modelling system may be integrated as part of the server 160. The modelling system 180 may include a model generator 182. The model generator 182 may be configured to generate a model. The generated model may be configured perform any one or more of: predict a fuel efficiency from measurement data, output the predicted fuel efficiency, isolate driver factors, isolate mechanical load factors, isolate mass load factors, determine a difference between the average fuel efficiency and the predicted fuel efficiency, quantify a contribution of each driver factor to the difference between the average fuel efficiency and the predicted fuel efficiency, and quantify a contribution of each mechanical load factor and / or mass load factor to the difference between the average fuel efficiency and the predicted fuel efficiency. Predicting the fuel efficiency may include determining the fuel efficiency by the generated model. The generated model may be stored in a model data 146 in the database 140. The model data 146 may include all information necessary for the generating and running the model, including any one or more of: a model architecture, model parameters, and model metadata. The modelling system may include a model environment 184. The model environment 184 may be configured to operate a generated model. Operating the generated model may include any one or more of: formatting input data for the modelling system, outputting the model results, providing an interface for interacting with the model, and providing a training process for training the model. The modelling system 180 may include a model training system 185. The model training system 185 may be configured to train a model using training data. The training data may be obtained from the database.

[0069] The server may include an analysis module 165. The analysis module 165 may be configured to perform an analysis of the obtained data and / or post-processed data. The analysis may include any one or more of: a clustering analysis; a classification; and, a trend analysis. The analysis may include an analysis of the actual fuel efficiency and / or the predicted fuel efficiency. The classification may include classifying the driver into different classifications. The classifications may be usable to determine the strengths and weaknesses of the driver. The classification may include ranking any one or more of: drivers, vehicles, vehicle fleets, and / or routes. The classification may be usable to provide automatic recommendations. The trend analysis may determine trends over time, driver or vehicle. The trend analysis may include a trend of actual and / or predicted vehicle fuel efficiency over defined periods of time.

[0070] The clustering and / or trend analysis may each be implemented using any one of: supervised learning, unsupervised learning, and reinforcement learning.

[0071] Alternatively, the analysis module may be performed by the modelling system 180. The server 160 may include a target determination module 166. The target determination module 166 may be configured to determine a target fuel efficiency for a driver.

[0072] The server may include a route planning module 168. The route planning module 168 may be configured to receive a planned route for a trip. The route planning module 168 may communicate with the modelling system. Alternatively, the route planning module may send the route to the modelling system. The modelling system may determine a predicted fuel efficiency for the route. The predicted fuel efficiency for the route may be dependent on the planned driver for the route.

[0073] The server may include an output module 167. The output module may be configured to output the analysis from the analysis module. Additionally, the output module may be configured to output a predicted fuel efficiency for a planned route. The output module may be configured to generate a graphical output of for display to the driver 108. The output module may be configured to output a signal to the vehicle to limit the RPM and / or the speed of the vehicle.

[0074] The driver 108 may have the driver device 110. The driver device 110 may include an identification component 112. The identification component 112 may include a unique driver identifier. The unique driver identifier may be configured to link a driver of the vehicle to an operation over the defined time period to the vehicle. The driver identifier may be a physical device, such as the driver device 110. The driver device may be a mobile phone of the driver. The driver device 110 may authenticate a driver via biometrics of the driver 108. The identification component 112 may be an application installed on the driver deice 110. The driver may log in to the application using biometric data or through a pin code. The identification component 112 may communicate with the driver authentication sensor 129 to authenticate the driver with the telemetry device. The driver device may communicate with the driver authentication sensor 129 via static Bluetooth low energy (BLE). Alternatively, the driver identifier may be a proximity tag. The driver 108 may bring the tag within a proximity of the driver authentication sensor 129 to authenticate the driver.

[0075] In an example of the driver identifier being an application on a mobile device, the driver authentication sensor 129 may perform a driver authentication through the server 160. In another example of the driver identifier being a proximity tag, the driver authentication sensor 129 may perform a driver authentication at the telemetry device 120.

[0076] The driver device 110 may include a vehicle fuel efficiency feedback application 114. The vehicle fuel efficiency feedback application 114 may be associated with the vehicle driver fuel efficiency feedback system 162. The vehicle fuel efficiency feedback application 114 may be configured to provide feedback to the driver. The feedback may be related to insights regarding driver factors influencing the fuel efficiency of one or more trips. Through the application 114, the driver may be shown a variance decomposition of areas for improvement to obtain a desired fuel efficiency of the trip. The vehicle driver fuel efficiency feedback application 114 may include a graphical interface separate to the mobile device through which feedback can be provided to the driver. The graphical interface may be a dashboard. The feedback application 114 may be configured in the form of a driver coaching and debriefing application. The vehicle fuel efficiency feedback application 114 may be configured to provide feedback related to a fuel efficiency of each vehicle. The feedback may be related to insights regarding mechanical load factors and / or mass load factors influencing the fuel efficiency of one or more trips.

[0077] During operation of the vehicle, the driver device may be required to remain within a proximity of the vehicle 102. If the device goes out of proximity, the vehicle may automatically shut down. Alternatively, the telemetry device 120 may indicate that the driver is no longer connected to the vehicle. The telemetry device 120 may communicate with a database 140. The database 140 may be located within the telemetry device or it may be an off-device storage facility. The database 140 may be a cloud-based database, or it may form part of a server 160. The database 140 may store any one or more of: a driver record 142, vehicle data 144, environmental data 145, derived data 147, model data 146, and training data 143. The driver record 142 may include any one or more of: a list of driver details, history of trips per driver, and driver authentication details.

[0078] The system 100 described above may implement a method for obtaining vehicle measurements and vehicle driver fuel efficiency determination and feedback. An example method 200 for obtaining vehicle measurements and vehicle driver fuel efficiency determination and feedback is illustrated in the swim-lane flow diagram of Figure 2, in which respective swim-lanes delineate steps, operations or procedures performed by respective entities or devices. The method may include steps performed at the telemetry device 120 and the server 160 or modelling system 180 (as shown in the figure).

[0079] Before determining a vehicle driver fuel efficiency for a particular trip, the modelling system may be configured to output the predicted fuel efficiency. Configuring the modelling system may include obtaining data, preparing the obtained data, and training a model usable for determining the vehicle driver fuel efficiency.

[0080] The server may obtain 280 a plurality of vehicle data and environmental data from previous operations of one or more vehicles over defined time periods. Obtaining 280 the plurality of vehicle data and environmental data may include receiving the plurality of data after each trip of the vehicle. In some examples, the plurality of data may be obtained from a plurality of vehicles over many trips of each vehicle. The server 160 and / or the modelling system 180 may prepare 282 a training dataset from the obtained vehicle data and environmental data. Preparing 282 the training dataset may include processing the dataset into a format suitable for training a model of the modelling system. The server 160 may train 284 the modelling system. Training 284 the modelling system may include training the model of the modelling system to output the predicted fuel efficiency to approximate an actual fuel efficiency over the defined time period from vehicle data.

[0081] The telemetry device 120 may authenticate 202 a driver. The authentication may include receiving a signal from a physical device unique to a driver, such as via a unique physical tag. Alternatively, the telemetry device may receive a BLE signal from a mobile device of the driver. In response to receiving the authentication signal, the telemetry device may link 204 the unique driver identifier to an operation of the vehicle. The operation of the vehicle may be an operation over the defined time period. During operation of the vehicle, the telemetry device may obtain 206 measurement data from the CAN bus and / or other sensors. Obtaining 206 measurement data may include obtaining vehicle data and environmental data. The measurement data may be obtained 206 and stored on the database in real-time, or once the operation of the vehicle is complete. The obtained measurement data may be data obtained over a defined time period.

[0082] The telemetry device may store 207 the measurement data on the telemetry device. Storing the measurement data on the telemetry device may include storing the data on a temporary storage. Storing the data on the telemetry device may include transmitting the data to the database in real time. Storing the data on the telemetry device may include formatting the data in the into a format suitable for transmission and / or storage.

[0083] The telemetry device may transmit 208 the measurement data from the telemetry device to the server. The server 160 may receive 240 the measurement data from the telemetry device. Transmitting the measurement data from the telemetry device may include deleting the measurement data from the telemetry device after transmission. Receiving 240 the measurement data may include storing the measurement data with the plurality of vehicle data and environmental data for use in training the modelling system.

[0084] The measurement data may be input 242 into the modelling system. The modelling system may be a modelling system trained 284 using the training dataset. Inputting 242 the measurement data into the modelling system may include starting the model environment with the model. Inputting 242 the data into the modelling system may include transmitting the input data to the model environment. The input data may be usable in the model environment.

[0085] If no average fuel efficiency value is available, the modelling system may determine 241 an average fuel efficiency value. Determining the average fuel efficiency may include accessing the vehicle and environmental data from the database 140. Determining the average fuel efficiency may include determining a variance around each data variable from the vehicle data and environmental data.

[0086] Modelling system may execute the model and output 244 the predicted fuel efficiency value. The measurement data and the predicted fuel efficiency value may be usable by the modelling system to isolate 246 driver factors. Isolating driver factors may include isolating driver factors from environmental factors. The environmental factors may be related to the predicted fuel efficiency. Each of the isolated driver factors may relate to the predicted fuel efficiency. Each driver factor may relate to a different aspect of a driver’s behaviour. The modelling system may isolate 247 mechanical load factors and / or mass load factors relating to the predicted fuel efficiency. Each of the mechanical load factors and each of the mass load factors may relate to the predicted fuel efficiency.

[0087] The modelling system may quantify 248 contributions of each driver factor. Quantifying the driver factors may include determining the difference between the average fuel efficiency value and the predicted fuel efficiency value. Quantifying the driver factors may include quantifying the contribution of each driver factor to the difference between the average fuel efficiency value and the predicted fuel efficiency value. Quantifying the contribution may include assigning a fuel efficiency contribution to different aspects of a driver’s behaviour. For example, quantifying the contribution may include assigning a fuel efficiency contribution to each driver factor. The contribution that each environmental factor has on the difference between the average and predicted fuel efficiency may be determined. The modelling system may quantify a contribution of the mechanical load factors to the difference between the average fuel efficiency and the predicted fuel efficiency. The modelling system may quantify a contribution of the weight load factors to the difference between the average fuel efficiency and the predicted fuel efficiency. The contributions of the mechanical load factors may be usable in performing a vehicle-to-vehicle comparison. The vehicle-to-vehicle comparison may provide insights into a required maintenance schedule of vehicles. Quantifying 248 contributions of each driver factor may include quantifying the contribution of the mechanical load factors and quantifying the contribution of the mass load factors.

[0088] The modelling system may determine 250 the target fuel efficiency. In an example, determining the target fuel efficiency may include subtracting the contributions of the driver factors from the predicted fuel efficiency value.

[0089] The server may perform 252 a driver analysis. The driver analysis may include performing any one or more of: a clustering analysis, a classification, and a trend analysis. The server may aggregate 253 vehicle data and environmental data across multiple time periods of previous vehicle operations. Aggregating 253 the data may include providing insights on any one or more of: drivers, vehicles, and a vehicle fleet. The vehicle fleet may include one or more vehicles.

[0090] The server may output 254 the analysis to the driver. The server may output 254 an analysis of driver factors over the defined time period. The server may output 254 an analysis of the assigned fuel efficiency contributions over the defined time period. Outputting the analysis may include transmitting the analysis to the driver device. Outputting the analysis may include transmitting the analysis to the dashboard and / or the application 114 on the driver device 110. Outputting the analysis may include outputting the analysis to a control display or dashboard of the operator or coordinator of one or more vehicles.

[0091] The server may output 255 a control signal to the vehicle. The control signal may be configured to be interpreted by hardware of the vehicle to limit an operational condition of the vehicle. In an example, the operational condition may be a speed of the vehicle. The control signal may be configured to limit the RPM of the engine or to reduce the vehicle speed. Outputting 255 the control signal may include providing the control signal to hardware of the vehicle.

[0092] In a practical example, the average fuel efficiency for a plurality of trips may be determined to be 2.0 km / L. After a trip of a driver has been completed, the actual fuel efficiency of that specific trip may have been 1.7 km / L, i.e. the trip was less efficient than the average. The difference is 0.3 km / L. The modelling system may determine that the driver factors contributed 0.2 km / L to the fuel efficiency difference, and environmental factors such as the road quality may have contributed 0.1 km / L to the fuel efficiency difference. Therefore, the target fuel efficiency for the driver should not be the average of 2.0 km / l, but should rather be 1.9km / L, as this may account for factors beyond a driver’s control.

[0093] Figure 3 is a flow diagram illustrating a method 300 for training the modelling system. The modelling system may include one or more models. The one or more models may include any one or more of: an artificial intelligence (Al) model, a machine learning (ML) model, a deep learning (DL) model, a Shapley decomposition model, a variance decomposition model. The deep learning model may include and neural network model, such as an artificial neural network.

[0094] The one or more models may be configured such that the output from a first model may be input into a second model. The first model may be an ML model. The second model may be a Shapley decomposition model. Alternatively, the second model may be another form of a variance decomposition model.

[0095] The ML model may be configured to output the predicted fuel efficiency in response to the measurement data being input into the model. In an example, the ML model may be an extreme gradient boosting (XGBoost) ML model. The ML model may be trained using training data obtained over multiple previous operations of the vehicle or vehicles. The ML model may receive, as input values, any one or more of: vehicle data, and environmental data.

[0096] The Shapley decomposition model may be configured to isolate driver and environmental factors that contribute to the difference between the average fuel efficiency and the predicted fuel efficiency. The Shapley decomposition may determine how much each variable of the vehicle data and environmental data contributed to the difference.

[0097] The Shapley decomposition may generate Shapley values. Shapley decomposition is a statistical method used to measure the contribution of individual predictors in a regression model. The individual predictors, according to aspects of this disclosure, are the variables of the vehicle data and environmental data. A common way of understanding a model is to examine coefficients for each variable. The coefficients may indicate how much the model output may change when each of the input variable change. While coefficients are suitable for indicating what will happen when the value of an input variable changes, they may not measure the overall importance of a variable. A Shapley decomposition is used to determine the importance of each variable.

[0098] The output from the Shapley decomposition may indicate that, for a specific route, the average fuel efficiency could not be attained, and that environmental factors would alter the fuel efficiency significantly irrespective of the manner of driving. This result may allow for a target fuel efficiency to be altered to account for the varying environmental factors. The combined contributions from the driver and environmental factors may sum up to the difference between the average fuel efficiency and the predicted fuel efficiency. The Shapley decomposition may receive, as input, the trained ML model and / or the input variables into the ML model.

[0099] It is appreciated that while a modelling system comprising a ML model and a Shapley decomposition is disclosed, any modelling system configured to output a predicted fuel efficiency, isolate driver factors, and quantify a contribution of each of the driver factors to a difference between an average fuel efficiency and the predicted fuel efficiency, may be used.

[0100] The one or more models may be configured to receive, as input, a planned route of a vehicle. The planned route may include an elevation profile including elevation data. The elevation profile may include information relating to elevation data of the planned route. The model may output a predicted fuel efficiency of the planned route.

[0101] The one or more models of the modelling system may be trained with training data 143. Training the modelling system may include obtaining raw incoming data 310. The raw incoming data may include a plurality of vehicle data 341 and / or environmental data 342 from previous operations of one or more vehicles over defined time periods. The vehicle data and environmental data may be timestamped data. The timestamped data may include data that is associated to a specific driver when the data was obtained and the time that the data was recorded, such that it is known which driver a specific set of data belongs to.

[0102] Deep learning may use artificial neural networks (ANNs). Examples of ANNs may include convolutional neural networks, or an arrangement of recurrent neural networks, such as long- short term memory (LSTM) networks suitable for time series applications. An ANN may consist of interconnected units, commonly referred to as neurons, as they are inspired by and resemble neurons of the brain. The units may consist of nodes and edges forming a connected network. ANNs may be configured in the form of a layered structure with an input at the first layer and an output at the final layer. The layers between the first and final layer may be hidden layers.

[0103] Each node in the ANN may receive a signal from one or more nodes in the preceding layer, starting from the input layer. The output of a node may be computed by an activation function, which may be a non-linear function of the sum of the inputs into each node in each layer. The output value of each node in the preceding layer is multiplied by a weighting value, which determines the strength of each nodes output value. Finally, the value that is determined at the final layer is the output of the ANN. For regression type ANNs, the output may contain only a single node with a value, or many nodes. Alternatively, classification type ANNs, the output may include multiple nodes, where each node provides the probability of a classification type. More complex ANNs architectures are better suited to specific tasks. In addition to the weights and activation functions of a regular ANN, a convolutional neural network (CNN) may apply a filter (or a kernel) onto a two-dimensional data structure to reduce the size of the hidden layers in the neural network, thereby reducing the number of weights within the neural network. A CNN is particularly suited to image-based tasks, where image data is often structured as a two- dimensional data structure.

[0104] Training the one or more models may be a computationally intensive and time consuming. The model environment may be hosted at a large computing infrastructure or a cloud computing infrastructure that can be accessed over a network. These resources may allow for dynamic computing resources to be dedicated to training the one or more models, after which the trained models can be downloaded to run on a separate application. In an example, the model may be trained and stored on the server.

[0105] The raw incoming data 310 may undergo a data preparation process 311 to separate components of the raw incoming data into various categories suitable for training, such as any one or more of: input data, output data, training data, verification data, and validation data. The data preparation process 311 may prepare 282 the training dataset from the obtained vehicle data and environmental data. The training data 143 may be input into a training process 313. The training process may be performed on a large computing cluster which may access the database 140 to obtain the training data when required. The training process 313 may include training 284 the modelling system. The training process 313 may result in a trained model 314, or multiple models as part of the modelling system, may be output. The trained machine learning model 314 may be stored as model data 146. In an example, the training process 313 may be a supervised learning training process, whereby the model parameters are adjusted such that the predicted fuel efficiency matches the actual fuel efficiency, as stored in the training data, is output from the model in response to vehicle data and / or environmental data being input into the model. It should be noted that training data is not limited to images and may include different types of input such as user entries and / or selections made via a user interface, scans and / or other input of textual information, and / or other training data.

[0106] The trained model may be usable in a runtime process 322. The runtime process may be executed in the modelling system or the server. When driver feedback is required, the vehicle data and / or environmental data may be formatted as input data 321. The input data 321 may be input into the runtime process 322. The runtime process may output 323 any one or more of: the predicted fuel efficiency, the driver factors contributions, the environmental factor contributions, the mechanical load factor, the mass load factor contributions, and target fuel efficiency. The output 323 may be usable in further downstream processes 324.

[0107] Various components may be provided for implementing the method described above with reference to Figure 2. Figure 4A and 4B are block diagrams illustrating example components which may be provided by a system for obtaining vehicle measurements and vehicle driver fuel efficiency determination and feedback.

[0108] The server 160 may include a processor 430 for executing the functions of components described below, which may be provided by hardware or by software units executing on the system. The software units may be stored in a memory component 432 and instructions may be provided to the processor 430 to carry out the functionality of the described components. In some cases, for example in a cloud computing implementation, software units arranged to manage and / or process data on behalf of the server 160 may be provided remotely. Some or all of the components may be provided by a software application downloadable onto and executable on the driver device 110.

[0109] Figure 4A illustrates components of the server 160. The server 160 may include a measurement obtaining component 402 arranged to obtain measurement data. The measurement obtaining component 402 may be arranged to obtain measurement data from any one or more of: the CAN bus, sensors on the telemetry device, and sensors in communication with the telemetry device.

[0110] The server 160 may include a modelling system interacting component 404 arranged to interact with the modelling system. Interacting with the modelling system may include transmitting data to the modelling system and receiving the output of the modelling system.

[0111] The server 160 may include a target determining component 406 arranged to determine a target fuel efficiency value. The target determining component 406 may receive results from the modelling system. In an example, the target determining component 406 may receive the predicted fuel efficiency and the fuel efficiency contributions of the driver factors.

[0112] The server 160 may include an analysis component 408 arranged to perform an analysis of the modelling system results. The analysis component 408 may be arranged to perform a clustering analysis and / or a trend analysis. The analysis component 408 may be arranged to classify the driver into different classifications.

[0113] The server 160 may include an outputting component 410 arranged to output results of the modelling system and postprocessing. Outputting the results may include storing the results in the database. Outputting the results may include outputting the analysis for further processing.

[0114] The server 160 may include a vehicle driver feedback component 412 arrange to provide feedback to the vehicle fuel efficiency feedback application. The vehicle driver feedback component 412 may configure the result for output to the vehicle fuel efficiency feedback application 114.

[0115] The server 160 may include an operator feedback component 414 arranged to provide feedback to an operator of one or more vehicles. The operator feedback component 414 may be arranged to provide aggregated feedback to a control centre of the operator of any one or more of: driver, drivers, vehicle, and vehicle fleet. The operator feedback component 414 may be arranged to provide feedback to a dashboard.

[0116] Figure 4B illustrates components of the modelling system 180. The modelling system may include a model environment component 450 arranged to setup and run an environment for the model. The model environment may include a virtual system arranged to perform any one or more of: running multiple models; configuring the models; training the models; saving the models to storage; transmitting the models; and the like. The modelling system may include a model generation component 452 arranged to generate one or more models. The model generation component 452 may interface with the database to store one or more generated models.

[0117] The modelling system may include a model training component 454 arranged to train one or more models. The model training component 452 may be arranged to receive training data from the database. The model training component 452 may be arranged to transmit a trained model to the database.

[0118] The modelling system may include a model operating component 456 arranged to perform functions of the modelling system. Such functions may include any one or more of: generate model; train model; store model; retrieve model; execute model; and the like.

[0119] The modelling system may include a data input component 458 arranged to provide input data to each of the one or more models. The data input component 458 may be configured to extract output data from one model for input into a subsequent downstream model.

[0120] The modelling system may include a prediction component 460 arranged to predict a fuel efficiency value for a set of input data. The prediction component may be arranged to determine the difference between the average fuel efficiency value and the predicted fuel efficiency.

[0121] The modelling system may include a driver factor isolation component 462 arranged to isolate driver factors. Isolating the driver factors includes isolating the factors from environmental factors.

[0122] The modelling system may include a vehicle factor isolation component 463 arranged to isolate mechanical load factors and / or mass load factors.

[0123] The modelling system may include a contribution quantification component 464 arranged to determine the contribution that each of the isolated driver factors have on the difference between the average value and the predicted fuel efficiency value. The contribution quantification component 464 may be arranged to determine the contribution that the mechanical load factors and / or the mass load factors have on the difference between the average value and the predicted fuel efficiency value.

[0124] Figure 5A illustrates an example graphical representation 500 for providing feedback according to aspects of the disclosure. The graphical representation 500 may be presented on a device within a vehicle. Alternatively, the graphical representation may be presented to the driver via the vehicle fuel efficiency feedback application 114. Additionally, the graphical representation 500 may be presented to the operator of one or more vehicles at an operator device or control centre.

[0125] The following display options may be selected to display results for any one of: all drivers, a subset of drivers, or per driver. All data presented to the driver may be selected across a date range 508 for a selected period. The graphical representation 500 may be configured to output a comparison of any one or more of: a driver-to-driver comparison, a vehicle-to-vehicle comparison, and a vehicle fleet to vehicle fleet comparison.

[0126] On a per driver level, the graphical representation 500 may display an aggregate of data from multiple trips for a selected driver. The graphical representation may display an overall efficiency score based on the modelling system output and / or analysis. The score may reflect how efficient the driver is across all trips. The graphical representation 500 may indicate which driver behaviour affects the fuel efficiency and where improvements may be made.

[0127] On a vehicle level, the graphical representation 500 may display the fuel efficiency per vehicle, irrespective of the driver. The vehicle data may allow operators to compare how different vehicles perform, taking into account mechanical factors.

[0128] On a vehicle fleet level, the graphical representation 500 may display the fuel efficiency for an entire fleet of vehicles, which may be usable to identify patterns or inefficiencies that may require attention.

[0129] The graphical representation 500 may include an average trip duration time 502. The display 500 may include a total distance travelled 504. The display 500 may include a trip count 506 of the total number of individual trips made. The price of fuel 509 may be output. The price of fuel 509 may be adjusted by a user. The fuel price may be displayed as a price per litre. The total number of assets 507 indicates the total number of vehicles being considered in the feedback presented. The total number of operators 505 may indicate the total number of drivers considered in the feedback.

[0130] The graphical representation 500 may include a total number of litres consumed 510. The actual fuel consumption 512, as obtained from the vehicle, may be used to derive the fuel efficiency. The fuel efficiency may be presented as the actual fuel consumption. Based on the results of the modelling system, the display may present the suggested or target fuel used 514. This is the total amount of fuel that would be used if the driver had altered driving behaviours according to feedback of the system, or when a signal to limit the RPM or speed of vehicle would be used. Additionally, a suggested or target fuel efficiency 516 may be presented. A predicted cost savings 518 may be presented. The cost savings 518 presented may be determined by a price of fuel and target reduction in fuel usage. A total potential quantity of fuel saved, a potential percentage of fuel saved, and / or a total monetary value of fuel saved may be presented. In the case of a nonfuel based vehicle, a potential energy saved may be presented.

[0131] The graphical representation 500 may present a breakdown of the contributing factors to the difference between the average fuel consumption and predicted fuel consumption. The breakdown may include the contributions for target fuel as a chart 520 with associated legend 522. The contributing factors may include any one or more of: driver, topography, and vehicle. In figure 5A, the driver contribution indicates that driver behaviour had the largest impact on fuel consumption, emphasising the need for driver training, feedback, and / or monitoring. The topography contribution indicates the influence of terrain and elevation change on fuel consumption. The vehicle contribution indicates the vehicle type, condition, or load on fuel efficiency. Furthermore, the graphical representation 500 may present a classification of trips according to distance or time per trip. The classification of trips may be provided as a pie chart 530 with an associated legend 532.

[0132] Figure 5B illustrates an example graphical representation 540 for providing a breakdown of the fuel savings per driver according to aspects of the disclosure. The display may present a bar chart which visualises the potential fuel savings per driver or vehicle, ranked from highest to lowest. It may identify the drivers with the largest potential for improving fuel efficiency. A legend indicates the measures 542 displayed. The measures may include a potential fuel savings in litres, provided as a bar 554 in the bar graph. The measures may include a potential savings in monetary value, provided as a circle 552 in the bar graph.

[0133] The potential fuel saving may be presented for any one or more of: potential savings by vehicle 550; a trip classification fuel consumption 548; a driver decomposition 546; and, a topography decomposition 544.

[0134] Figure 5C illustrates an example graphical representation 560 for providing a trend analysis of the fuel savings per driver according to aspects of the disclosure. The trend analysis may be a time-series bar chart displaying a potential fuel savings in litres over time. The analysis may highlight monthly trends. A bar chart displaying a potential fuel savings in litres and monetary value may be shown for a time A 568 and a time B 569 separately. An estimated amount of fuel in litres that could be saved may be shown by the line 570. A line chart may indicate a comparison between the vehicle fleet’s actual weekly fuel efficiency 564 and the target fuel efficiency 562. This comparison may allow the vehicle operator to see where their performance stands relative to best practices in industry. The time period may be adjusted to daily, weekly, monthly, or any other suitable time frame.

[0135] Figure 6 illustrates an example of a computing device 600 in which various aspects of the disclosure may be implemented. The computing device 600 may be embodied as any form of data processing device including a personal computing device (e.g. laptop or desktop computer), a server computer (which may be self-contained, physically distributed over a number of locations), a client computer, or a communication device, such as a mobile phone (e.g. cellular telephone), satellite phone, tablet computer, personal digital assistant or the like. Different embodiments of the computing device may dictate the inclusion or exclusion of various components or subsystems described below.

[0136] The computing device 600 may be suitable for storing and executing computer program code. The various participants and elements in the previously described system diagrams may use any suitable number of subsystems or components of the computing device 600 to facilitate the functions described herein. The computing device 600 may include subsystems or components interconnected via a communication infrastructure 605 (for example, a communications bus, a network, etc.). The computing device 600 may include one or more processors 610 and at least one memory component in the form of computer-readable media. The one or more processors 610 may include one or more of: CPUs, graphical processing units (GPUs), microprocessors, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs) and the like. In some configurations, a number of processors may be provided and may be arranged to carry out calculations simultaneously. In some implementations various subsystems or components of the computing device 600 may be distributed over a number of physical locations (e.g. in a distributed, cluster or cloud-based computing configuration) and appropriate software units may be arranged to manage and / or process data on behalf of remote devices.

[0137] The memory components may include system memory 615, which may include read only memory (ROM) and random access memory (RAM). A basic input / output system (BIOS) may be stored in ROM. System software may be stored in the system memory 615 including operating system software. The memory components may also include secondary memory 620. The secondary memory 620 may include a fixed disk 621 , such as a hard disk drive, and, optionally, one or more storage interfaces 622 for interfacing with storage components 623, such as removable storage components (e.g. magnetic tape, optical disk, flash memory drive, external hard drive, removable memory chip, etc.), network attached storage components (e.g. NAS drives), remote storage components (e.g. cloud-based storage) or the like.

[0138] The computing device 600 may include an external communications interface 630 for operation of the computing device 600 in a networked environment enabling transfer of data between multiple computing devices 600 and / or the Internet. Data transferred via the external communications interface 630 may be in the form of signals, which may be electronic, electromagnetic, optical, radio, or other types of signal. The external communications interface 630 may enable communication of data between the computing device 600 and other computing devices including servers and external storage facilities. Web services may be accessible by and / or from the computing device 600 via the communications interface 630.

[0139] The external communications interface 630 may be configured for connection to wireless communication channels (e.g., a cellular telephone network, wireless local area network (e.g. using Wi-Fi™), satellite-phone network, Satellite Internet Network, etc.) and may include an associated wireless transfer element, such as an antenna and associated circuitry. The external communications interface 630 may include a subscriber identity module (SIM) in the form of an integrated circuit that stores an international mobile subscriber identity and the related key used to identify and authenticate a subscriber using the computing device 600. One or more subscriber identity modules may be removable from or embedded in the computing device 600.

[0140] The external communications interface 630 may further include a contactless element 650, which is typically implemented in the form of a semiconductor chip (or other data storage element) with an associated wireless transfer element, such as an antenna. The contactless element 650 may be associated with (e.g., embedded within) the computing device 600 and data or control instructions transmitted via a cellular network may be applied to the contactless element 650 by means of a contactless element interface (not shown). The contactless element interface may function to permit the exchange of data and / or control instructions between computing device circuitry (and hence the cellular network) and the contactless element 650. The contactless element 650 may be capable of transferring and receiving data using a near field communications capability (or near field communications medium) typically in accordance with a standardized protocol or data transfer mechanism (e.g., ISO 14443 / NFC). Near field communications capability may include a short-range communications capability, such as radio-frequency identification (RFID), Bluetooth™, infra-red, or other data transfer capability that can be used to exchange data between the computing device 600 and an interrogation device. Thus, the computing device 600 may be capable of communicating and transferring data and / or control instructions via both a cellular network and near field communications capability. The computer-readable media in the form of the various memory components may provide storage of computer-executable instructions, data structures, program modules, software units and other data. A computer program product may be provided by a computer-readable medium having stored computer-readable program code executable by the central processor 610. A computer program product may be provided by a non-transient or non-transitory computer- readable medium, or may be provided via a signal or other transient or transitory means via the communications interface 630.

[0141] Interconnection via the communication infrastructure 605 allows the one or more processors 610 to communicate with each subsystem or component and to control the execution of instructions from the memory components, as well as the exchange of information between subsystems or components. Peripherals (such as printers, scanners, cameras, or the like) and input / output (I / O) devices (such as a mouse, touchpad, keyboard, microphone, touch-sensitive display, input buttons, speakers and the like) may couple to or be integrally formed with the computing device 600 either directly or via an I / O controller 635. One or more displays 645 (which may be touch- sensitive displays) may be coupled to or integrally formed with the computing device 600 via a display or video adapter 640.

[0142] The computing device 600 may include a geographical location element 655 which is arranged to determine the geographical location of the computing device 600. The geographical location element 655 may for example be implemented by way of a global positioning system (GPS), or similar, receiver module. In some implementations the geographical location element 655 may implement an indoor positioning system, using for example communication channels such as cellular telephone or Wi-Fi™ networks and / or beacons (e.g. Bluetooth™ Low Energy (BLE) beacons, iBeacons™, etc.) to determine or approximate the geographical location of the computing device 600. In some implementations, the geographical location element 655 may implement inertial navigation to track and determine the geographical location of the communication device using an initial set point and inertial measurement data.

[0143] The foregoing description has been presented for the purpose of illustration; it is not intended to be exhaustive or to limit the technology to the precise forms disclosed. Persons skilled in the relevant art can appreciate that many modifications and variations are possible in light of the above disclosure.

[0144] Any of the steps, operations, components or processes described herein may be performed or implemented with one or more hardware or software units, alone or in combination with other devices. Components or devices configured or arranged to perform described functions or operations may be so arranged or configured through computer-implemented instructions which implement or carry out the described functions, algorithms, or methods. The computer- implemented instructions may be provided by hardware or software units. In one embodiment, a software unit is implemented with a computer program product comprising a non-transient or non- transitory computer-readable medium containing computer program code, which can be executed by a processor for performing any or all of the steps, operations, or processes described. Software units or functions described in this application may be implemented as computer program code using any suitable computer language such as, for example, Java™, C++, or Perl™ using, for example, conventional or object-oriented techniques. The computer program code may be stored as a series of instructions, or commands on a non-transitory computer-readable medium, such as a random access memory (RAM), a read-only memory (ROM), a magnetic medium such as a hard-drive, or an optical medium such as a CD-ROM. Any such computer-readable medium may also reside on or within a single computational apparatus, and may be present on or within different computational apparatuses within a system or network.

[0145] Flowchart illustrations and block diagrams of methods, systems, and computer program products according to embodiments are used herein. Each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, may provide functions which may be implemented by computer readable program instructions. In some alternative implementations, the functions identified by the blocks may take place in a different order to that shown in the flowchart illustrations.

[0146] Some portions of this description describe the examples in terms of algorithms and symbolic representations of operations on information. These algorithmic descriptions and representations, such as accompanying flow diagrams, are commonly used by those skilled in the data processing arts to convey the substance of their work effectively to others skilled in the art. These operations, while described functionally, computationally, or logically, are understood to be implemented by computer programs or equivalent electrical circuits, microcode, or the like. The described operations may be embodied in software, firmware, hardware, or any combinations thereof.

[0147] 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. It is therefore intended that the scope of the present disclosure be limited not by this detailed description, but rather by any claims that issue on an application based hereon. Accordingly, the present disclosure is intended to be illustrative, but not limiting, of the scope of any accompanying claims. Finally, throughout the specification and any accompanying claims, unless the context requires otherwise, the word ‘comprise’ or variations such as ‘comprises’ or ‘comprising’ will be understood to imply the inclusion of a stated integer or group of integers but not the exclusion of any other integer or group of integers.

Claims

CLAIMS:

1. A computer-implemented method for vehicle (102) driver fuel efficiency feedback, comprising: obtaining (206) measurement data of vehicle data (144) and environmental data (145) during operation of the vehicle (102) over a defined time period; inputting (242) the measurement data into a modelling system (180) configured to: output (244) a predicted fuel efficiency from the measurement data; isolate (246) driver factors relating to the predicted fuel efficiency from environmental factors related to the predicted fuel efficiency, wherein each driver factor relates to a different aspect of a driver’s behaviour; and, quantify (248) a contribution of each driver factor to a difference between an average fuel efficiency and the predicted fuel efficiency, wherein quantifying includes assigning a fuel efficiency contribution to each driver factor; and, outputting (254) an analysis of the driver factors and the assigned fuel efficiency contributions over the defined time period, wherein the average fuel efficiency is an average fuel efficiency value determined using measurement data obtained from previous operations of one or more vehicles over defined time periods.

2. The method as claimed in claim 1 , including isolating (247) mechanical load factors and mass load factors relating to the predicted fuel efficiency; and, quantifying a contribution of the mechanical load factors and quantifying a contribution of the mass load factors to the difference between the average fuel efficiency and the predicted fuel efficiency.

3. The method as claimed in claims 1 or 2, including: obtaining (280) a plurality of vehicle data and environmental data from previous operations of one or more vehicles over defined time periods; preparing (282) a training dataset from the obtained vehicle data and environmental data; and, training (284) the modelling system to output the predicted fuel efficiency to approximate an actual fuel efficiency over the defined time period from vehicle data.

4. The method as claimed in any one of the preceding claims, including determining (250) a target fuel efficiency by subtracting the contributions of each driver factor from the predicted fuel efficiency value.

5. The method as claimed in any one of the preceding claims, including aggregating (253) vehicle (144) and environmental data (145) across multiple time periods to provide insights, wherein the insights include information for any one or more of: drivers, vehicles, and vehicle fleets.

6. The method as claimed in any one of the preceding claims, wherein the modelling system (180) is configured to perform (252) a clustering analysis and / or a trend analysis of an actual fuel efficiency and predicted fuel efficiency values.

7. The method as claimed in any one of the preceding claims, wherein outputting (244) the predicted fuel efficiency includes determining a distribution of the fuel efficiency around the predicted fuel efficiency value.

8. The method as claimed in any one of the preceding claims, including providing a telemetry device (120) associated with the vehicle and configured to obtain (206) the vehicle data and environmental data.

9. The method as claimed in claim 8, wherein the telemetry device (120) obtains (206) the vehicle data from a controller area network (CAN) bus (104) of the vehicle (102).

10. The method as claimed in claim 8 or 9, wherein the telemetry device (120) is configured to link a driver (108) of the vehicle to an operation of the vehicle via a driver identifier, wherein the driver identifier is a physical device configured to uniquely identify and authenticate a driver (108) of one or more drivers.

11. The method as claimed in any one of the preceding claims, wherein the environmental data (145) include any one or more of: global positioning system (GPS) coordinates, elevation data, road type, and road quality.

12. The method as claimed in any one of the preceding claims, wherein the vehicle data (144) includes any one or more of: actual fuel efficiency, vehicle speed, vehicle acceleration, vehicle coasting, and vehicle engine revolutions per minute (RPM), fuel efficiency, torque, kilowatt output, gross combination mass, vehicle make, and vehicle type.

13. The method as claimed in any one of the preceding claims, wherein the vehicle data (144) and the environmental data (145) include a timestamp associated with each measurement.

14. The method as claimed in any one of the preceding claims, wherein isolating (246) the driver factors relating to fuel efficiency includes isolating (246) a set of pre-defined driver factors, and wherein the driver factors include any one or more of: RPM management, speed management, and acceleration management, driver consistency, or any other operator classification.

15. The method as claimed in any one of the preceding claims, wherein quantifying (248) the contribution includes quantifying (248) the contribution of each environmental factor to the difference between the average fuel efficiency and the predicted fuel efficiency.

16. The method as claimed in any one of the preceding claims, including providing a graphical output representation (500) of comparison information, wherein the comparison information includes any one or more of: driver-to-driver comparison, vehicle-to-vehicle comparison, vehicle fleet to vehicle fleet comparison.

17. The method as claimed in any one of the preceding claims, including providing a signal to limit the engine RPM or limiting the vehicle speed in response to specific measurement data.

18. The method as claimed in any one of the preceding claims, wherein the modelling system (180) is configured to receive, as input, a planned route for a vehicle and output a predicted fuel efficiency for the planned route, and wherein the planned route includes an elevation profile of the planned route.

19. The method as claimed in any one of the preceding claims, wherein the modelling system (180) includes any one or more of: an artificial intelligence model, a machine learning model, a deep learning model, and a Shapley decomposition model.

20. The method as claimed in any one of the preceding claims, wherein the modelling system (180) includes a machine learning model or a neural network model, and is configured to receive as input the vehicle data (144) and / or environmental data (145) and output a predicted fuel efficiency value.

21. The method as claimed in any one of the preceding claims, wherein the modelling system (180) includes a Shapley decomposition model, configured to receive any one or more of: a model (314) configured to output a predicted fuel efficiency; vehicle data; and environmental data, and configured to output isolated driver factors relating to the predicted fuel efficiency and to quantify the contribution of each driver factor to the difference between the average fuel efficiency and the predicted fuel efficiency.

22. A system for vehicle (102) driver fuel efficiency feedback, conducted at a server, comprising: a non-transitory computer-readable storage medium; and one or more processors coupled to the non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium comprises program instructions that, when executed on the one or more processors, cause the system to perform operations comprising: obtaining (206) measurement data of vehicle data (144) and environmental data (145) during operation of the vehicle (102) over a defined time period; inputting (242) the measurement data into a modelling system (180) configured to: output (244) a predicted fuel efficiency from the measurement data; isolate (246) driver factors relating to the predicted fuel efficiency from environmental factors related to the predicted fuel efficiency, wherein each driver factor relates to a different aspect of a driver’s behaviour; and, quantify (248) a contribution of each driver factor to a difference between an average fuel efficiency and the predicted fuel efficiency, wherein quantifying includes assigning a fuel efficiency contribution to each driver factor; and, outputting (254) an analysis of the driver factors and the assigned fuel efficiency contributions over the defined time period, wherein the average fuel efficiency is an average fuel efficiency value determined using measurement data obtained from previous operations of one or more vehicles over defined time periods.

23. The system as claimed in claim 22, including the modelling system (180), wherein the modelling system (180) includes a machine learning model, configured to receive vehicle data (144) and / or environmental data (145), and configured to output a predicted fuel efficiency value, and wherein the modelling system (180) includes a Shapley decomposition model, configured to receive vehicle data (144) and / or environmental data (145), and configured to output isolated driver factors relating to the predicted fuel efficiency and to quantify the contribution of each driver factor to the difference between the average and predicted fuel efficiency24. A computer program product for determining a fuel efficiency of a vehicle (102), comprising: a non-transitory computer-readable storage medium; and one or more processors coupled to the non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium comprises program instructions that, when executed on the one or more processors, cause the system to perform operations comprising: obtaining (206) measurement data of vehicle data (144) and environmental data (145) during operation of the vehicle (102) over a defined time period; inputting (242) the measurement data into a modelling system (180) configured to: output (244) a predicted fuel efficiency from the measurement data; isolate (246) driver factors relating to the predicted fuel efficiency from environmental factors related to the predicted fuel efficiency, wherein each driver factor relates to a different aspect of a driver’s behaviour; and, quantify (248) a contribution of each driver factor to a difference between an average fuel efficiency and the predicted fuel efficiency, wherein quantifying includes assigning a fuel efficiency contribution to each driver factor; and, outputting (254) an analysis of the driver factors and the assigned fuel efficiency contributions over the defined time period, wherein the average fuel efficiency is an average fuel efficiency value determined using measurement data obtained from previous operations of one or more vehicles over defined time periods.

Images