ADAPTIVE DIAGNOSTIC PARAMETRISATION

By grouping vehicles and adjusting leak detection classifiers based on ground truth data, the system improves the accuracy of fuel tank leak detection, addressing false positives and negatives in existing systems.

DE112017007416B4Active Publication Date: 2024-11-14FORD GLOBAL TECH LLC
View PDF 1 Cites 0 Cited by

Patent Information

Application Number
DE112017007416
Authority / Receiving Office
DE · DE
Patent Type
Patents
Current Assignee / Owner
Filing Date
2017-05-05
Publication Date
2024-11-14
Estimated Expiration
2037-05-05

AI Technical Summary

Technical Problem

Existing leak detection systems in vehicles struggle to accurately detect small fuel leaks in fuel tanks, often resulting in false positives or negatives due to varying environmental and operational conditions.

Method used

A system that groups vehicles based on specification and environmental data, adjusts leak detection classifiers using ground truth data, and applies tuned classifiers to prevent false detections by considering vehicle-specific and environmental factors.

Benefits of technology

Enhances the accuracy of fuel tank leak detection by minimizing false positives and negatives through adaptive classification, ensuring reliable leak detection across varying conditions.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure 00000000_0000_ABST
    Figure 00000000_0000_ABST
Patent Text Reader

Abstract

System comprising a processor programmed to do the following: Defining a large number of vehicle groups based on vehicle specification data; Defining a multitude of subgroups for each of the vehicle groups based on environmental and sensor data received from each of a multitude of vehicles, encompassing the definition of subgroups of a selected vehicle group by: Determining a statistical correlation of the vehicle sensor data and environmental data assigned to the vehicles included in the selected vehicle group, and Identifying two or more vehicles from the vehicle group based on a specific statistical correlation and a deviation threshold; and Setting fuel tank leak detection classifiers for the subgroups based on ground truth data, including, for each of the multitude of vehicles, a leak detection status and a leak test result.
Need to check novelty before this filing date? Find Prior Art

Description

STATE OF THE ART

[0001] Vehicles can be powered by internal combustion engines. A vehicle with an internal combustion engine typically has a fuel tank that provides fuel, such as gasoline, to the engine. A fuel tank can develop a leak, resulting in fuel escaping from the fuel tank. Due to the chemical properties of engine fuels, a fuel leak can cause environmental problems. Therefore, vehicles are typically equipped with leak detection systems to detect a leak in the fuel tank. Typically, a leak detection system is expected to detect very small fuel leaks, i.e., a hole in the fuel tank with a diameter of or greater than 0.02 inches (0.508 mm). Detecting such small leaks can be challenging for the leak detection system. BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a block diagram of an example vehicle system. Fig. 2 is a diagram showing a fuel tank and an engine of the vehicle from Fig. 1, which is fluidly connected via a fuel hose. Fig. 3A-3C are example diagrams illustrating a reference classifier and adjusted classifiers for leak detection in a fuel tank. Fig. 4 is a flowchart of an exemplary process for setting the reference classifier for leak detection in a fuel tank. Fig. Figure 5 is a flowchart of an example process for leak detection in a fuel tank. DETAILED DESCRIPTION INTRODUCTION

[0002] Disclosed herein is a system that includes a processor programmed to define a plurality of vehicle groups based on vehicle specification data. The processor is programmed to define a plurality of subgroups for each of the vehicle groups based on environmental data and sensor data received from each of a plurality of vehicles. The processor is further programmed to adjust fuel tank leak detection classifiers for the subgroups based on ground truth data, including, for each of the plurality of vehicles, a leak detection status and a leak test result.

[0003] The vehicle specification data may include one or more of vehicle model, fuel tank volume, fuel tank thermal characteristics, engine volume, position of heat-dissipating vehicle components relative to the fuel tank, transmission characteristics, fuel tank shape, exhaust gas parameters, and fuel tank pressure sensor position relative to the fuel tank.

[0004] The vehicle sensor data may include one or more of engine operating status, fuel tank pressure, fuel tank temperature, engine coolant temperature, fuel parameters, vehicle orientation, vehicle location coordinates, fuel gauge reading, last trip duration, last trip conditions, and last trip route.

[0005] The environmental data may include one or more of air temperature, air pressure, precipitation status, and wind conditions.

[0006] The processor may be further programmed to define vehicle groups by identifying two or more vehicles that share at least one identical parameter included in the vehicle specification data.

[0007] The processor may be further programmed to define subgroups of a selected vehicle group, determine a statistical correlation of the vehicle sensor data and environmental data associated with the vehicles included in the selected vehicle group, and identify two or more vehicles from the vehicle group based on the determined statistical correlation and a deviation threshold.

[0008] The statistical correlation of the vehicle sensor data may include a geographical distribution of the vehicles in the selected vehicle group.

[0009] The deviation threshold may include a maximum dimension threshold associated with an area in which the vehicles of the subgroup are geographically located.

[0010] The processor may be further programmed to adjust the fuel tank leak detection classifiers by identifying a leak detection curve threshold based on the vehicle sensor data and the ground truth data.

[0011] The ground truth data may further include vehicle sensor data and environmental data from a time of leak detection in the fuel tank.

[0012] The leak detection curve threshold can be determined based on a data classification method.

[0013] The processor may be further programmed to identify a falsely detected leak in a vehicle of the plurality of vehicles, wherein a leak detection status of the vehicle indicates a detected leak and a leak test result of the vehicle is negative, and adjust the classifiers for the subset including the vehicle based on the falsely detected leak.

[0014] The leak detection status may further include a time of leak detection and vehicle operating data from the time of leak detection.

[0015] The processor may be further programmed to perform leak detection based on the set classifiers.

[0016] The leak test result may be a smoke test result.

[0017] Further disclosed herein is a method including defining a plurality of vehicle groups based on vehicle specification data, defining a plurality of subgroups for each of the vehicle groups based on environmental data and sensor data received from each of a plurality of vehicles, and setting fuel tank leak detection classifiers for the subgroups based on ground truth data including, for each of the plurality of vehicles, a leak detection status and a leak test result.

[0018] The vehicle specification data may include one or more of vehicle model, fuel tank volume, fuel tank thermal characteristics, engine volume, position of heat-dissipating vehicle components relative to the fuel tank, transmission characteristics, fuel tank shape, exhaust gas parameters, and fuel tank pressure sensor position relative to the fuel tank.

[0019] The vehicle sensor data may include one or more of engine operating status, fuel tank pressure, fuel tank temperature, engine coolant temperature, fuel parameters, vehicle orientation, vehicle location coordinates, fuel gauge reading, last trip duration, last trip conditions, and last trip route.

[0020] Defining vehicle groups may further include identifying two or more vehicles that share at least one identical parameter included in the vehicle specification data.

[0021] Defining subgroups of a selected vehicle group may further include determining a statistical correlation of the vehicle sensor data and environmental data associated with the vehicles included in the selected vehicle group and identifying two or more vehicles from the vehicle group based on the determined statistical correlation and a deviation threshold.

[0022] Further disclosed is a computing device programmed to perform any of the above method steps. Furthermore, a vehicle including the computing device is disclosed.

[0023] Furthermore, a computer program product is disclosed that comprises a computer-readable medium storing instructions executable by a computer processor to perform any of the above method steps. EXAMPLE SYSTEM ELEMENTS

[0024] Fig. 1-2 illustrate a vehicle 100. The vehicle 100 may be powered in various known ways, for example, with a powertrain including an internal combustion engine 170. The vehicle 100 may include a hybrid powertrain, for example, including a combination of an internal combustion engine 170 and an electric motor. Although depicted as a passenger vehicle, the vehicle 100 may be any automobile, such as a car, a truck, an SUV, a crossover vehicle, a van, a minivan, etc. The vehicle 100 may include a computer 110, actuator(s) 120, sensor(s) 130, and a human-machine interface (HMI) 140.

[0025] The vehicle 100 includes one or more fuel tanks 150 that store fuel, e.g., gasoline, for operating the engine 170 of the vehicle 100. The fuel tank 150 may be fluidly coupled to the engine 170 via a fuel hose 220. The vehicle 100 may have a fuel fill inlet 160 for refilling the fuel tank 150. The fuel fill inlet 160 may be fluidly coupled to the fuel tank 150. The vehicle 100 may have a fuel cap 180 that closes the fuel fill inlet 160. A user may fill the fuel tank 150, e.g., at a gas station, by removing (e.g., unscrewing) the fuel cap 180 and tightening (e.g., screwing) the fuel cap 180 after filling the fuel tank 150. The fuel tank 150 can be arranged at any location in the vehicle 100 at a distance d1 from the engine 170 (see Fig. 2). For example, the fuel tank 150 may be located in a rear portion of the vehicle 100, and the engine 170 may be located in a front portion of the vehicle. The fuel tank 150 and the engine 170 may be coupled via a fuel hose 220, through which fluid (liquid) fuel is supplied from the tank 150 to the engine 170. The fuel tank 150 may be made of, for example, steel, plastic, etc. For example, a plastic fuel tank 150 may be formed from high-density polyethylene.

[0026] Computer 110 includes a processor and memory, as known in the art. Memory includes one or more forms of computer-readable media and stores instructions executable by computer 110 for performing various operations, including those disclosed herein.

[0027] The computer 110 may include programming to control one or more of the following: vehicle braking, propulsion (e.g., controlling acceleration within the vehicle by controlling one or more of an internal combustion engine, electric motor, hybrid engine, etc.), steering, air conditioning, interior and / or exterior lighting, etc.

[0028] Computer 110 may include or be communicatively coupled to more than one processor, e.g., controllers or the like included in vehicle 100 for monitoring and / or controlling various controls of vehicle 100, e.g., a powertrain controller, a braking controller, a steering controller, etc. Computer 110 is generally arranged for communication on a communication network of vehicle 100, which may include a Controller Area Network (CAN) or the like and / or other wired and / or wireless mechanisms.

[0029] Through the network of the vehicle 100, the computer 110 may transmit messages to and / or receive messages from various devices in the vehicle 100, e.g., an actuator 120, an HMI 140, etc. Alternatively or additionally, in cases where the computer 110 actually includes multiple devices, the vehicle communications network may be used for communications between devices, which are represented in this disclosure as the computer 110.

[0030] The actuators 120 of the vehicle 100 are implemented via circuits, chips, or other electronic and / or mechanical components that can actuate various vehicle subsystems according to the respective control signals, as is known. The actuators 120 can be used to control vehicle systems such as braking, acceleration, and / or steering of the vehicle 100.

[0031] Additionally, the computer may be configured to communicate with a remote computer 180 via a network 190 via a vehicle-to-infrastructure (V-to-I) interface. Network 190 represents one or more mechanisms by which computer 110 and remote computer 180 may communicate with each other, and may be one or more different wired or wireless communication mechanisms, including any combination of wired (e.g., cable and fiber) and / or wireless (e.g., cellular, wireless, satellite, microwave, and radio frequency) communication mechanisms and any desired network topology (or topologies if multiple communication mechanisms are used). Example communication networks include wireless communication networks (e.g., using one or more cellular networks, Bluetooth, IEEE 802.11, etc.).), dedicated short-range communications (DSRC), local area networks (LAN) and / or wide area networks (WAN), including the Internet, that provide data communication services.

[0032] The HMI 140 may be configured to receive user inputs, for example, during operation of the vehicle 100. Furthermore, an HMI 140 may be configured to present information to the user. Thus, the HMI 140 may be located in a passenger compartment of the vehicle 100. For example, the HMI 140 may provide information to the user, including a detected leak in the fuel tank 140. As explained below, a leak may be detected after the engine 170 is turned off, when most electrical components are off; therefore, the HMI 140 may first provide information, including a detected leak, in the next drive cycle, e.g., the next time the engine 170 is turned on.

[0033] The sensors 130 of the vehicle 100 may include a variety of known devices for providing sensor 130 data over the communication bus of the vehicle 100. The sensor 130 data of the vehicle 100 may include engine operating status, e.g., running, off, etc., fuel tank 140 pressure, fuel tank 150 temperature, engine coolant temperature 170, fuel parameters (e.g., Reid vapor pressure), orientation of the vehicle 100, location coordinates of the vehicle 100, fuel gauge reading, last trip duration, last trip conditions, and last trip route.

[0034] For example, sensors 130 may include temperature and / or pressure sensors 130 disposed within vehicle 100. A pressure and / or temperature sensor 130 may provide temperature and / or pressure data encompassing an exterior (e.g., ambient) region of vehicle 100. A pressure and / or temperature sensor 130 may include a transducer 130 that converts an applied pressure and / or temperature, for example, into a change in an electrical parameter such as electrical resistance, capacitance, etc.

[0035] In another example, as in Fig. 2, a pressure and / or temperature sensor 130 may provide data including an interior of the fuel hose 220, i.e., a temperature and / or pressure within the hose 220. The sensor 130 may be mounted to the fuel hose 220 at a distance d2, e.g., 10 cm from the fuel tank 150. The distance d2 may refer to a length of a portion of the hose 220 disposed between the sensor 130 and the outer surface of the fuel tank 150. A pressure within the fuel tank 150 may be related to the pressure at the sensor 130 mounted on the hose 220. Therefore, the computer 110 may be programmed to determine a temperature and / or pressure within the fuel tank 150 based at least in part on the received data from the sensor 130, the distance d2, and / or a diameter d4 of a fuel hose 220.Additionally or alternatively, a sensor 130 may be mounted on the fuel tank 150 that determines the pressure and / or temperature inside the fuel tank 150.

[0036] The sensors 130 may include a fuel gauge sensor 130 that determines a fuel level remaining in the fuel tank 150. The computer 110 may be programmed to determine a volume of air and / or fuel vapor remaining in the fuel tank 150 based on data obtained from the fuel gauge sensor 130. For example, the computer 110 may be programmed to determine a volume of air and / or fuel vapor based on a current level of fuel in the tank 150, and the known capacity of the fuel tank 150, e.g., 50 liters, may be stored in the memory of the computer 110.

[0037] The vehicle 100 may include a global positioning system (GPS) sensor 130 configured to determine coordinates of a current location of the vehicle 100. The computer 110 may be programmed to transmit the location coordinates of the vehicle 100 to the remote computer 185 via the network 190. In one example, the remote computer 185 may be programmed to determine environmental data of the vehicle 100 based at least in part on the determined location coordinates of the vehicle 100. Environmental data includes data associated with an area (or vicinity) around the vehicle 100, e.g., an area with a 1 kilometer (km) radius around the vehicle 100. The environmental data of the vehicle 100 may include barometric pressure, temperature, wind conditions, precipitation status, etc., in an area around the vehicle 100.In one example, the remote computer 185 may be programmed to determine the environmental data based on information received from weather stations distributed at various geographic locations. Additionally or alternatively, the computer 185 may be programmed to receive environmental data from sensors 130 of the vehicle 100, such as an ambient air temperature and / or pressure sensor 130 located in the bumper of the vehicle 100.

[0038] The fuel of vehicle 100 is typically flammable and may contain chemical contaminants that are hazardous to the environment, e.g., soil, groundwater, etc. Thus, a fuel leak from fuel tank 150 may pose an environmental and / or safety concern. Fuel may leak from fuel tank 150 in the form of a liquid fluid and / or a vapor. For example, fuel cap 180 may not be properly closed and / or left open after refueling vehicle 100. In another example, a hole 210 may be present in the fuel tank. Hole 210 may be located anywhere on fuel tank 150 or other components, such as fuel inlet hose 160, that are fluidly connected to fuel tank 150.In one example, the computer 110 of the vehicle 100 may be programmed to determine whether there is a leak in the vehicle fuel tank 150 (or other components such as the hose 220, the inlet 160, etc.) that has a diameter d3 greater than or equal to 0.02 inches (0.508 mm). A leak hole 210 may have a rectangular, oval, circular, or asymmetrical shape, such as a crack.

[0039] Exposing the vehicle 100 to different environmental conditions, such as different ambient air pressures, different driving conditions, etc., may result in incorrect detection of a leak in the fuel tank 150. For example, the computer 110 of the vehicle 100 may incorrectly detect a leak when no leak actually exists (also referred to herein as a false positive); and / or the computer 110 may fail to detect an existing leak in the fuel tank 150 (also referred to herein as a misdetection).

[0040] With reference to Fig. 1-2, the remote computer 185 may be programmed to define one or more vehicle groups based on vehicle 100 specification data, described in more detail below, and to define one or more subgroups for each of the vehicle groups 100 based on the environmental data and the sensor 130 data received from each of a plurality of vehicles 100. The remote computer 185 may be further programmed to set fuel tank 150 leak detection classifiers for the subgroups based on ground truth data, including, for each of the vehicles 100, a leak detection status and a leak test result. Each of the computers 110 of the vehicles 100 may be programmed to receive the set classifiers from the remote computer 185 and perform leak detection based on the set classifiers.Thus, adjusting the classifier may advantageously prevent false detection and / or misidentification of a leak in a fuel tank 150 of the vehicle 100. In one example, the remote computer 185 may be programmed to provide the adjusted classifiers to the computers 110 only after confirmation of the adjusted classifiers by a third-party computer, e.g., a standards authority.

[0041] The computer 110 of the vehicle 100 may be programmed to operate leak detection, i.e., evaluate whether a leak exists in the fuel tank 150, when it is determined that certain diagnostic entry conditions are met. For example, the computer 110 may be programmed to determine whether the engine 170 of the vehicle 100 is off based on the received status data of the engine 170. Additionally or alternatively, the data from the sensors 130 may include the speed of the vehicle 100, etc. The computer 110 may be programmed to determine whether the vehicle 100 is stationary based on the received speed of the vehicle 100. The computer 110 may be programmed to determine that the entry conditions are met when it is determined that the engine 170 is off.Additionally, the computer 110 may be programmed to determine that the entry conditions are met when it is determined that the fuel tank 150 is not being refueled. The computer 110 may be programmed to operate the leak detection for a predetermined period of time, e.g., 30 minutes, after the engine 170 is shut down. In one example, the computer 110 may operate in a low-power mode to reduce the power consumption of the computer 110 since the engine 170 is shut down in the leak detection state.

[0042] The specification data of the vehicle 100 may include one or more of a model of the vehicle 100, a type of the engine 170 (e.g., gasoline, hybrid, etc.), a volume of the fuel tank 150, thermal properties of the fuel tank 150 such as thermal resistance, material of the fuel tank 150, displacement of the engine 170, position of the heat-dissipating vehicle components such as the engine 170, batteries, etc. relative to the fuel tank 150 (e.g., distance d1 in Fig. 2), transmission type of the vehicle 100 (four-wheel drive, all-wheel drive, etc.), shape of the fuel tank 150, exhaust configuration (e.g., heat emission parameters of the exhaust gas), whether there is a fuel tank cap or the tank is capless, and position of the pressure sensor 130 of the fuel tank 150 relative to the fuel tank 150 (e.g., the distance d2 in Fig. 2). A transmission type of the vehicle 100 may affect the density of the components of the vehicle 100 to one another, i.e., how close they are arranged to one another. A packing density of components of the vehicle 100 may affect the amount of heat to which the tank 150 is exposed. An increase in heat typically increases pressure inside the tank 150, which affects leak detection. As another example, an exhaust configuration may have an impact on the heat dissipation of the fuel tank 150. The fuel inlet 160 may have various mechanisms for closing, for example, with the cap 180 or a capless mechanism as is now often used. As discussed below, the vehicles 100 may be grouped based on the specification data of the vehicle 100 to have vehicles 100 in a group that are suitable for comparison with one another for leak detection conditions.

[0043] The leak detection status of each of the vehicles 100 may include a leak condition and a no-leak condition. In one example, the leak condition may mean that a leak with a diameter equal to or greater than 0.02 inches is present. The no-leak condition may mean either that there is no leak or that the diameter of the leak is less than 0.02 inches. The computer 110 of the vehicle 100 may be programmed to determine the leak detection status based on the leak detection classifiers and the data from the sensor 130 of the vehicle 100. Additionally, the leak detection status may include an inconclusive condition, meaning the computer 110 cannot conclude whether a leak is present or not. For example, the classifier may include a hysteresis range that is 1% less and / or greater than a threshold at which no decision can be made.The computer 110 of each of the vehicles 100 may be programmed to store the leak detection status in a memory of the computer 110 and / or transmit the leak detection status to the remote computer 185. The computer 110 of the vehicle 100 may be further programmed to output data to the HMI 140 of the vehicle 100 based on the leak detection status. For example, the HMI 140 may include a malfunction indicator lamp (MIL), and the computer 110 may be programmed to illuminate the malfunction indicator lamp when a leak is detected in the fuel tank 150 of the vehicle 100.

[0044] A fuel tank 150 may be tested, e.g., in a service facility, a laboratory, etc., to determine whether a leak exists in the fuel tank 150. A result of such a test is referred to herein as a leak test result, while a determination made by the computer 110 that a leak is present or not is referred to herein as a leak detection status. For example, the fuel tank 150 may be tested using a smoke testing technique to determine whether a leak exists in the tank 150. In a smoke test, a chemical in the fuel tank 150 produces visible smoke, which, when seen outside the tank 150, may indicate a leak in the tank 150. The vehicle 100 may be tested in the workshop to verify whether a detected leak was a false positive or a correctly detected leak. Similarly, if no leak is detected by the computer 110 of the vehicle 100, the leak test result may determine a leak, i.e.,the computer 110 of the vehicle 100 has correctly determined that there is no leak in the tank 150.

[0045] The ground truth data associated with each vehicle 100 includes the leak test result and the leak detection status of the respective vehicle 100. As stated above, the leak detection status indicates whether or not a leak is detected (determined by the computer 110) and may further include a time and date for determining the leak. In other words, the ground truth data for each vehicle 100 indicates whether a detection of a leak and / or a determination that no leak is present was correct. The ground truth data indicates a correct detection of a leak when the leak detection status and the leak test result are the same, i.e., both determine a leak or both determine no leak. The ground truth data may indicate an incorrect detection of a leak when the leak detection status and the leak test result differ, e.g.,the leak detection status indicates a leak, but the leak test result concludes that there is no leak, and / or the leak detection status indicates no leak, whereas the leak test result concludes that there is a leak in the tank 150. As described below with respect to the . Fig. 3A, the classifier can be defined based on ground truth data. As discussed with reference to Fig. As further explained in Figures 3B-3C, the classifiers may be adjusted based on new ground truth data, e.g., after computer 185 receives new ground truth data indicating a false positive.

[0046] In one example, the ground truth data may be determined based on the date and time of the leak detection status. In one example, the computer 110 of the vehicle 100 may store operational data of the vehicle 100 (i.e., data available from sensors and / or controls of the vehicle 100, e.g., via a vehicle network such as a communication bus) at the time a leak is determined, e.g., temperature and / or pressure in the fuel tank 150 and / or the ambient temperature around the vehicle 100, the last driving conditions of the engine 170, etc. Additionally, the computer 110 may be programmed to store the operational data periodically, e.g., every hour, even if it is concluded that there is no leak in the tank 150. As described below with respect to Fig. 3C, this information may be necessary to determine the operational data associated with a false detection of a leak being present (i.e., a time at which the computer 110 incorrectly concluded that no leak was present).

[0047] The remote computer 185 may be further programmed to receive the operating data of the vehicle 100 at the time of determining the leak detection status and determine a set classifier based on the stored information. In one example, the computer 110 of the vehicle 100 may have determined a leak while the vehicle was in an elevated area, e.g., a mountainous area. However, the leak test result of the tank 150 of the vehicle 100 may indicate a false-positive detection (i.e., no leak). The computer 185 may Fbased on operating data in the mountainous area rather than based on current operating data, e.g., in a service center at low altitude and at different pressure and / or temperature. Advantageously, stored operating data can provide insight into the conditions under which the false-positive detection occurred.

[0048] Defining vehicle groups 100 may further include identifying two or more vehicles 100 that share at least one identical parameter, e.g., the volume of the fuel tank 150, which data is typically included in the specification data of the vehicle 100. The computer 185 may be programmed to receive the vehicle 100 specification data from the vehicle 100. The computer 185 may be programmed to define the vehicle groups 100 based on one or more parameters included in the specification data of the vehicle 100. In one example, the computer 185 may be programmed to determine that vehicles 100 of the same model are included in a group.In another example, the computer 185 may be programmed to determine that vehicles 100 with the same thermal properties of the fuel tank 150, the same displacement of the engine 170, and the same distance d1 between the engine 170 and the fuel tank 150 are included in the same group. The fuel tanks 150 of the vehicles 100 in the same group may have similar thermodynamic properties, e.g., a similar volume, similar thermal conductivity, a similar heat source, etc., and therefore, the ground truth data from one vehicle 100 in a group may be relevant to another vehicle 100 in the same group. For example, ground truth data showing a false positive at a particular operating time in one of the vehicles 100 in a group may be relevant to identifying a classifier setting applicable to all vehicles 100 in that group.

[0049] On the other hand, in addition to parameters from the specification data of the vehicle 100, other parameters at a time of leak detection, such as environmental data, e.g., ambient air pressure and / or temperature, data from the sensor 130 of the vehicle 100, such as current driving conditions, including the temperature of the engine 170, etc., may influence the leak detection results and may therefore be taken into account. In other words, groups of vehicles 100 grouped based on specification data of the vehicle 100 may be further subdivided (as discussed below) to define subgroups based on dynamically changing vehicle data (compared to typically immutable specification data of the vehicle 100), such as data from the sensor 130 of the vehicles 100, environmental data, etc., to identify the vehicles 100 in subgroups.

[0050] In one example, defining subgroups of a vehicle group may further include determining a statistical correlation of the data of the sensor 130 of the vehicle 100 and the environmental data associated with the vehicles 100 included in the selected vehicle group, and identifying two or more vehicles 100 from the vehicle group 100 based on the determined statistical correlation and a deviation threshold.

[0051] The statistical correlation of the vehicle sensor data may include a geographical distribution of the vehicles 100 of a respective group of vehicles 100. The deviation threshold may include a maximum dimensional threshold associated with an area in which the vehicles 100 of the subgroup are geographically located. For example, vehicles 100 of a vehicle group 100 that are geographically located in a circular area with a radius of 5 km may be classified into a subgroup. The vehicles 100 of a subgroup of vehicles 100 may have an engine 170 temperature at a shutdown time of the engine 170 that is within a certain range, e.g., 500 to 600 degrees Celsius.In yet another example, the computer 185 may be programmed to receive the temperature of the engine 170 from the vehicles 100 via the network 190 and identify two or more vehicles 100 from a respective group of vehicles 100 based on the received temperature data, e.g., based on the average temperature of 520 degrees Celsius and a deviation of 15 degrees Celsius.

[0052] The leak detection classifiers of the fuel tank 150 (also referred to as classifiers for simplicity) classify a fuel tank 150 of the vehicle 100 as a leaking and a non-leaking fuel tank 150. A classifier can be a linear or a non-linear line (see Fig. 3A). Computer 110 may be programmed to determine whether a leak exists in tank 150 based on the position of the operating condition relative to the classifier, e.g., above or below a threshold. In the context of this disclosure, an "operating condition" is a combination of two or more parameters of vehicle 100 at a particular time during operation of vehicle 100, where the parameters include data from sensor 130 and possibly other data such as environmental data, etc. For example, classifier C classifies Ref , as in Fig. 3A, operating states, which are a combination of temperature and pressure of the fuel tank 150, are classified as leaking and non-leaking. The operating point of the vehicle 100 can be assigned to a leaking or non-leaking fuel tank 150 based on the ground truth data. The operating states "O" represent exemplary operating states of a (correctly detected) leaking fuel tank 150, while the operating states "X" represent exemplary (correctly detected) non-leaking fuel tanks included in a subset. In other words, the classifier C classifies Ref the operating states “O” “X” of a subgroup of vehicles 100.

[0053] In one example, measurements are performed in a laboratory to identify the operating conditions "O" for the respective detected leaks and the operating conditions "X" for the respective times at which a leak was not detected. For example, a valve with an orifice larger than 0.02 inches may be mounted on the tank, and the operating conditions of the tank are determined to represent the operating conditions "O." Similarly, operating conditions of a non-leaking tank 150 are measured in the laboratory to identify several non-leaking points "X." Then, using a mathematical method such as a data classification method, the classifier C Refthat separates all leaking operating states "O" from all non-leaking operating states "X". In one example, the operating states include more than two parameters, e.g., temperature and pressure in the tank 150, temperature of the engine 170, ambient temperature, etc. Thus, the classifier can be a multidimensional plane, surface, etc. Data classification methods can include support vector machine (SVM), relevance vector machine (RVM), etc.

[0054] As explained above, the computer 185 may be programmed to adjust the classifiers to prevent further false positives and / or misdetection of a leak. In one example, the computer 185 may be programmed to identify a falsely detected leak in a particular vehicle 100 using data from a subset of vehicles and based on the leak detection status of the vehicle 100 indicating a detected leak and a leak test result of the vehicle 100 indicating no leak. The computer 185 may be programmed to adjust the classifiers for the subset including the vehicle 100 with the falsely detected leak.In other words, based on the ground truth indicating a false detection in a subgroup of vehicle 100, computer 185 can determine a set classifier that can be applied to other vehicles 100 in the same subgroup. A classifier set in this way can result in such a false detection being avoided in the future by computer 110 of other vehicles 100 in this subgroup.

[0055] In one example, a group of vehicles 100 includes vehicles with a fuel tank volume 150 of 50 and a specific engine type 170. The remote computer 185 may identify a number, e.g., 10, of vehicles 100 at a specific altitude with a specific ambient air pressure as a subgroup. The remote computer 185 may then adjust a classifier based on the ground truth data for this subgroup. For example, the computer 185 may, as in Fig. 3A, based on the received ground truth data, determine that the operating state O F contains a false positive, ie a leaking fuel tank 150 is incorrectly classified based on the classifier C Ref identified as detected. The computer 185 may be programmed to identify a set leak detection classifier, e.g., a curve threshold C1, based on the sensor data of the vehicle 100 and the ground truth.

[0056] In other words, the remote computer 185 may identify a classifier C1 that (i) classifies the previously identified non-leaking operating conditions “X” and the identified false detection operating condition O Ffrom (ii) the previously identified leaky operating conditions "O". The computer 185 can then transmit data including the adjusted classifier to vehicles 100 (not limited to the vehicles 100 identified as a subset of vehicles, as mentioned above). Thus, any other computer 110 of the vehicle 100 that receives the adjusted classifier can prevent such a false detection when the operating condition O F of the vehicle 100, for example, the operating state O F Thus, if another vehicle 100 of the above-mentioned group, which was not identified in the subgroup, joins this subgroup, e.g., if the user operates the vehicle to drive to a height associated with the above-mentioned subgroup, then the set classifier C1 can prevent false detection of a leak.

[0057] In another example, which is Fig. 3C, the computer 185 may receive ground truth data, including a leak detection status appropriate for operating state X F indicates no leak, whereas a laboratory leak test result indicates a leak in the tank 150. The computer 185 may be programmed to identify a set leak detection classifier, e.g., a curve threshold C2, based on the sensor data of the vehicle 100 and the ground truth. In other words, the remote computer 185 may identify a classifier that distinguishes (i) the previously identified non-leaking operating conditions "X" from (ii) the identified operating condition X Fand the previously identified leaking operating conditions "O." Computer 185 can then transmit data including the adjusted classifier C2 to vehicles 100 (not limited to the vehicles 100 identified as a subset of vehicles, as mentioned above). Thus, any other computer 110 of vehicle 100 receiving the adjusted classifier C2 can detect the leak in operating condition X. F In other words, if another vehicle 100 of the above-mentioned group, which was not identified in the sub-group, joins this sub-group, e.g., if the user operates the vehicle to drive to such a height that is assigned to the above-mentioned sub-group, then the set classifier C2 can correctly detect the leak, e.g., the computer 100 can be programmed to detect a leak in operating state X F based on the set classifier C2.

[0058] As stated above, an operating condition can be multidimensional and therefore a classifier can be multidimensional, e.g., a surface. Thus, a tuned classifier can include a tuned surface. For example, a tuning of a classifier surface can result in a dent, a protrusion, etc., being generated on a surface based on received ground truth. For example, a classifier surface can have a dent to classify an operating condition differently based on an identified misdetection or false positive. In other words, a surface classifier, as in tuning curve C Ref in the Fig. 3B-3C to set an operating state X F or O Fto classify differently based on the received ground truth data, have a dent or protrusion to classify an operating condition differently to prevent misdetection or false positive. PROCESSING

[0059] Fig. 4 is a flow diagram of an example process 400 for setting a leak detection classifier to detect a leak of a fuel tank 150 of a vehicle 100. For example, the remote computer 185 may be programmed to execute blocks of the process 400.

[0060] The process 400 begins at a block 410 where the remote computer 185 receives data from vehicles 100, other computers such as weather station computers, etc. The received data may include specification data of the vehicles 100, environmental data, data from sensors 130 of the vehicles 100, ground truth data, etc. For example, the computer 185 may receive data from sensors 130 of the vehicle 100, including temperature and / or pressure of the fuel tank 150 of the vehicles 100. The computer 185 may receive ground truth data from another remote computer, such as a manufacturer's dealer computer.

[0061] Next, in a block 420, the remote computer 185 defines groups of vehicles 100, e.g., based on specification data of the vehicle 100. For example, the remote computer may be programmed to determine two groups of vehicles: (i) vehicles 100 with a displacement of 2.2 liters in the engine 170 and a 50-liter fuel tank 150, and (ii) vehicles 100 with a displacement of 3.6 liters in the engine 170 and a 70-liter fuel tank 150. The remote computer 185 may be programmed to define groups based on any other combination of parameters from the specification data of the vehicles 100.

[0062] Next, in a block 430, the remote computer 185 defines one or more subgroups of vehicles 100 for each of the defined vehicle groups 100. The remote computer 185 may define subgroups based on operational data of the vehicle 100, including data from sensors 140 of the vehicles 100, environmental data associated with each of the vehicles 100, etc. In one example, the remote computer 185 may be programmed to define subgroups based on geographic distributions of vehicles 100 in each group. For example, a subgroup of the vehicle group 100 "displacement of 2.2 liters in the engine 170 and 50-liter fuel tank 150" may include the vehicle 100 at a certain altitude range (e.g., above sea level) and / or a typical ambient barometric pressure range (which may be identified based on data from the GPS sensor 130 of the vehicle 100).

[0063] Next, in a decision block 440, the remote computer 185 determines whether adjustment of one or more classifiers is warranted. For example, the remote computer 185 may be programmed to determine that adjustment of the classifier(s) is warranted after identifying a false positive (detection of a nonexistent leak) and / or a false detection (failure to detect a present leak) based on received ground truth data. If the remote computer 185 determines that adjustment of the classifier(s) is warranted, the process 400 proceeds to block 450; otherwise, the process 400 ends or, alternatively, returns to block 410, although this is not indicated in Fig. 4 is not shown.

[0064] In block 450, the remote computer 185 determines a set leak detection classifier. For example, the remote computer 185 may be programmed to identify a set classifier, e.g., the curve thresholds C1, C2, such as a Fig. 3B-3C, based on data from sensor 130 of vehicle 100 and the ground truth data.

[0065] Next, in a block 460, the remote computer 185 transmits the adjusted classifier(s) to vehicles 100, e.g., via a network 190. In one example, the remote computer 185 may be programmed to transmit a message to vehicles 100, including notification of a setting of leak detection classifiers. The remote computer 185 may be further programmed to transmit data, including the adjusted classifier(s), to a vehicle 100 after receiving a request for the adjusted classifier(s). As another example, the remote computer 185 may be programmed to send the adjusted classifier data to the vehicle 100 periodically, e.g., every second.

[0066] After block 460, the process 400 ends or alternatively returns to block 410, although this is Fig. 4 is not shown.

[0067] Fig. 5 is a flow diagram of an example process 500 for detecting a leak in a fuel tank 150 of a vehicle 100. For example, a computer 110 of the vehicle 100 may be programmed to execute blocks of the process 500.

[0068] The process 500 begins at a block 510, where the computer 110 of the vehicle 100 receives data. For example, the computer 110 of the vehicle 100 may be programmed to receive data from sensor 130, e.g., pressure and / or temperature of the fuel tank 150, status of the engine 170, etc. The computer 110 may further be programmed to receive environmental data, e.g., ambient air temperature and / or pressure, precipitation status, etc.

[0069] Next, in a decision block 520, the computer 110 determines whether the diagnostic entry condition(s) are met. For example, the computer 110 may be programmed to determine whether the vehicle 100 is stationary and the engine 170 of the vehicle 100 is in a turned-off state. In another example, the computer 110 may be programmed to determine that the diagnostic conditions are met if it is determined that the engine 170 is in the turned-off state and the time elapsed since the engine 170 was turned off is less than 30 minutes. If the computer 110 determines that the diagnostic entry condition(s) are met, then the process 500 continues to a decision block 530; otherwise, the process 500 returns to block 510.

[0070] In decision block 530, computer 110 determines whether a preset classifier is available. In one example, computer 110 may receive a message from remote computer 185 and determine whether the message includes notification of a classifier change. In another example, computer 110 may be programmed to determine whether preset classifier data is being broadcast from remote computer 185. In yet another example, the computer may be programmed to transmit a message to remote computer 185 including a request for preset classifiers. Additionally or alternatively, at a service center for vehicle 100, such as an auto dealership, service personnel may connect a programming device to the network of vehicle 100, and the programming device may store the preset classifiers in computer 110 of vehicle 100.The programming device may receive the set classifiers from the remote computer 185. If the computer 110 determines that the set classifiers are available, the process 500 continues to a block 540; otherwise, the process 500 continues to a decision block 550.

[0071] In block 540, the computer 110 stores the adjusted classifiers in the memory of the computer 110. In one example, the computer may be programmed to store the received adjusted classifiers in a non-volatile memory of the computer 110.

[0072] In decision block 550, the computer 110 determines whether a leak is present in the fuel tank. The computer 110 may be programmed to determine whether a leak with a diameter greater than or equal to 0.02 inches is present based on the leak detection classifier. The leak detection classifier(s) may be reference classifiers, e.g., classifier C Ref and / or a set classifier, e.g., classifier(s) C1, C2 (depending on whether decision block 550 is reached from decision block 530 or block 540). If the computer 110 determines that a leak is present in the fuel tank 150, the process 500 continues to a block 560; otherwise, the process 500 ends or, alternatively, returns to block 510, although this is not Fig. 5 is not shown.

[0073] Next, in block 560, the computer 110 performs an operation based on the detected leak. For example, the computer 110 may be programmed to actuate the HMI 140 of the vehicle 100 to illuminate an HMI 140 malfunction indicator light. After block 560, the process 500 ends or, alternatively, returns to block 510, although this is not Fig. 5 is not shown.

[0074] The article "a / e / r" used to modify a noun should be understood to mean one or more, unless otherwise specified or the context indicates otherwise. The expression "based on" includes partially or wholly based on.

[0075] Computing devices, as discussed herein, generally include instructions executable by one or more computing devices such as those identified above, and for performing blocks or steps of processes described above. Computer-executable instructions may be compiled or interpreted by computer programs created using a variety of programming languages ​​and / or technologies, including, among others, either alone or in combination, Java™, C, C++, Visual Basic, Java Script, Perl, HTML, etc. In general, a processor (e.g., a microprocessor) receives instructions, e.g., from memory, a computer-readable medium, etc., and executes those instructions, thereby performing one or more processes, including one or more of the processes described herein.Such instructions and other data may be stored and transmitted using a variety of computer-readable media. A file in the computing device is generally a collection of data stored on a computer-readable medium, such as a storage medium, random access memory, etc.

[0076] A computer-readable medium includes any medium involved in providing data (e.g., instructions) that can be read by a computer. Such a medium can take many forms, including, but not limited to, non-volatile media, volatile media, and so on. Non-volatile media includes, for example, optical or magnetic disks and other persistent storage. Volatile media includes dynamic random access memory (DRAM), which is typically a main memory.Common forms of computer-readable media include, for example, a floppy disk, a film disk, a hard disk, a magnetic tape, any other magnetic medium, a CD-ROM, a DVD, any other optical medium, punched cards, punched tape, any other physical medium with hole patterns, a RAM, a PROM, an EPROM, a FLASH EEPROM, any other memory chip or any other memory cartridge, or any other medium that can be read by a computer.

[0077] With respect to the media, processes, systems, methods, etc. described herein, it is understood that although the steps of such processes, etc., have been described as occurring according to a particular sequence, such processes could be performed with the described steps performed in a different order than the order described herein. It is further understood that certain steps could be performed concurrently, other steps could be added, or certain steps described herein could be omitted. In other words, the descriptions of systems and / or processes in this specification are provided for the purpose of illustrating particular embodiments and should in no way be construed to limit the disclosed subject matter.

[0078] Accordingly, it is to be understood that the present disclosure, including the foregoing description and the appended figures and following claims, is intended to be illustrative and not restrictive. Many embodiments and applications other than the examples provided will become apparent to those skilled in the art upon reading the foregoing description. The scope of the invention should be determined not with reference to the foregoing description, but instead with reference to claims appended hereto and / or included in a non-provisional patent application based hereon, along with the full scope of equivalents to which such claims are entitled.It is anticipated and intended that there will be future developments in the techniques discussed herein, and that the disclosed systems and methods will be incorporated into such future embodiments. It is understood that the disclosed subject matter is susceptible to modification and variation.

Claims

[1] System comprising a processor programmed to: Defining a variety of vehicle groups based on vehicle specification data; Defining a plurality of subgroups for each of the vehicle groups based on environmental data and sensor data received from each of a plurality of vehicles, comprising defining subgroups of a selected vehicle group by: Determining a statistical correlation of the vehicle sensor data and environmental data associated with the vehicles included in the selected vehicle group, and Identifying two or more vehicles from the vehicle group based on the determined statistical correlation and a deviation threshold; and Setting fuel tank leak detection classifiers for the subgroups based on ground truth data including, for each of the plurality of vehicles, a leak detection status and a leak test result. [2] The system of claim 1, wherein the vehicle specification data includes one or more of vehicle model, fuel tank volume, thermal characteristics of the fuel tank, engine volume, position of heat-dissipating vehicle components relative to the fuel tank, transmission characteristics, shape of the fuel tank, exhaust gas parameters, and position of the fuel tank pressure sensor relative to the fuel tank. [3] The system of claim 1, wherein the vehicle sensor data includes one or more of engine operating status, fuel tank pressure, fuel tank temperature, engine coolant temperature, fuel parameters, vehicle orientation, vehicle location coordinates, fuel gauge reading, last trip duration, last trip conditions, and last trip route. [4] The system of claim 1, wherein the environmental data includes one or more of air temperature, air pressure, precipitation status, and wind conditions. [5] The system of claim 1, wherein the processor is further programmed to define vehicle groups by identifying two or more vehicles that share at least one identical parameter included in the vehicle specification data. [6] The system of claim 1, wherein the statistical correlation of the vehicle sensor data includes a geographical distribution of the vehicles of the selected vehicle group. [7] The system of claim 6, wherein the deviation threshold includes a maximum dimension threshold associated with an area in which the vehicles of the subgroup are geographically located. [8] The system of claim 1, wherein the processor is further programmed to adjust the fuel tank leak detection classifiers by identifying a leak detection curve threshold based on the vehicle sensor data and the ground truth data. [9] The system of claim 8, wherein the ground truth data further includes vehicle sensor data and environmental data from a time of leak detection in the fuel tank. [10] The system of claim 8, wherein the leak detection curve threshold is determined based on a data classification technique. [11] The system of claim 8, wherein the processor is further programmed to: Identifying at least one falsely detected leak in a vehicle of the plurality of vehicles, wherein a leak detection status of the vehicle indicates a detected leak and a leak test result of the vehicle is negative; and Adjusting the classifiers for the subgroup containing the vehicle based on the falsely detected leak. [12] The system of claim 1, wherein the leak detection status further includes a time of leak detection and vehicle operating data from the time of leak detection. [13] The system of claim 1, wherein the processor is further programmed to perform leak detection based on the set classifiers. [14] The system of claim 1, wherein the leak test result is a smoke test result. [15] Procedure comprising: Defining a variety of vehicle groups based on vehicle specification data; Defining a plurality of subgroups for each of the vehicle groups based on environmental data and sensor data received from each of a plurality of vehicles, comprising defining subgroups of a selected vehicle group by: Determining a statistical correlation of the vehicle sensor data and environmental data associated with the vehicles included in the selected vehicle group, and Identifying two or more vehicles from the vehicle group based on the determined statistical correlation and a deviation threshold; and Setting fuel tank leak detection classifiers for the subgroups based on ground truth data including, for each of the plurality of vehicles, a leak detection status and a leak test result. [16] The method of claim 15, wherein the vehicle specification data includes one or more of vehicle model, fuel tank volume, thermal characteristics of the fuel tank, engine volume, position of heat-dissipating vehicle components relative to the fuel tank, transmission characteristics, shape of the fuel tank, exhaust gas parameters, and position of the fuel tank pressure sensor relative to the fuel tank. [17] The method of claim 15, wherein the vehicle sensor data includes one or more of engine operating status, fuel tank pressure, fuel tank temperature, engine coolant temperature, fuel parameters, vehicle orientation, vehicle location coordinates, fuel gauge reading, last trip duration, last trip conditions, and last trip route. [18] The method of claim 15, wherein defining vehicle groups further includes identifying two or more vehicles that share at least one identical parameter included in the vehicle specification data.

Citation Information

Patent Citations

  • Vehicle diagnostic or prognostic message transmission systems and methods

    US8036788B2