Apparatus and method for testing automated vehicles

Abstract

The processor is responsive to a set of location or motion data describing one or more objects relative to the first local frame of reference to generate a transformed set of location or motion data describing the one or more objects relative to a second local frame of reference different from the first local frame of reference such that the set of location or motion data and the transformed set of location or motion data relative to the global frame of reference are identical, and outputs the transformed set of location or motion data to the vehicle such that the vehicle performs a control action in response thereto.

Description

[Technical Field]

[0001] Technical field This application claims the interests of U.S. Provisional Patent Application No. 62 / 984,683, filed on 3 March 2020, and all disclosures of that patent document are incorporated herein by reference.

[0002] This disclosure relates to the calibration of autonomous vehicle control systems and the measurement of their effects on the energy efficiency of a host vehicle and on exhaust gas emissions of internal combustion engine vehicles having autonomous vehicle control systems. More specifically, this application relates to characterizing real-world gas emissions from internal combustion engine (ICE) vehicles, including hybrid electric vehicles (HEVs), and real-world energy efficiency and automatic braking response of any vehicle type having autonomous longitudinal speed control or braking features, including battery electric vehicles (BEVs), based on laboratory tests. [Background technology]

[0003] background Modern automobiles can reliably operate under almost every combination of environmental conditions, road gradients, and driving conditions found on Earth. Such vehicles are common throughout the world and operate uniformly and reliably in ambient temperatures ranging well below 0°C to over 40°C, from dry deserts to humid rainforests, and from congested, slow-speed urban traffic to high-speed operation on the German autobahn.

[0004] Many countries with large automobile populations have exhaust emission or vehicle efficiency standards that automobile manufacturers must comply with. However, experience has shown that testing vehicles under a wide range of real-world environments, roads, and driving conditions known to affect vehicle emissions, fuel consumption, or energy efficiency in the real world is difficult and costly. Furthermore, it is well known that the energy efficiency of HEVs and the mileage of BEVs per charge decrease at relatively low ambient temperatures.

[0005] Laboratory-based tailpipe emission and energy efficiency tests have so far been conducted under limited ambient conditions, vehicle speed schedules, and driving conditions. Given the dramatic increase in the number of vehicles worldwide in recent years, and the increasing computerization of vehicles, it is necessary for governments and automakers to have a relatively good understanding of vehicle emissions across a relatively wide range of operating conditions, so that National Environmental Air Quality (NAAQ) standards can continue to be met within the "realizable area" and ultimately be met within the current "non-realizable area." It is also necessary for vehicle manufacturers to be able to evaluate the effects of potential changes to vehicle emission control and powertrain calibration across a relatively wide range of ambient and operating conditions.

[0006] The difficulties associated with studying and understanding the "real-world" emission of vehicles are further complicated in the case of autonomous and semi-autonomous vehicles, such as those with longitudinal velocity control or adaptive cruise control (ACC). This is because, while laboratory dynamometers provide realistic real-world loads to the vehicle, it is not yet known how to make a vehicle with longitudinal velocity control operate in such a way that the longitudinal velocity control autonomously controls its speed in the same manner as it controls speed in the real world. Being able to operate a vehicle in this manner in the laboratory is particularly useful for evaluating the effects of changes made to longitudinal velocity control calibration or vehicle powertrain calibration, and is the subject of discussion herein.

[0007] Laboratory test methods are known to be very accurate and repeatable for emissions and efficiency measurements under actual test conditions, but real-world driving can expose vehicles to a relatively wide range of conditions for which conventional laboratory test protocols are not. This is due to a number of factors including the difficulty of simulating the full range of real-world temperature and atmospheric pressure conditions in the laboratory, the influence of real-world driver behavior under actual traffic conditions, etc.

[0008] At present, it is not possible to "blind test" a vehicle that is autonomously controlled within a laboratory environment in order to measure exhaust emissions, fuel consumption, or vehicle efficiency while the vehicle is communicating autonomously with other vehicles (a blind test requires that no changes be allowed to the subject vehicle except for the change being evaluated, and that no detailed technical knowledge of the vehicle control system is required). However, autonomous features are likely to affect emissions and energy efficiency in most vehicles, particularly if the vehicle powertrain changes to utilize different calibrations compared to the calibrations that would be utilized for similar real-world driving under conventional driver control.

[0009] As the depth of technical knowledge possessed by vehicle and autonomous system developers increases, it may be possible to isolate and simulate the effects that other competing vehicles in real-world traffic have on automated vehicle control systems in order to gain some confidence in the system's operation in the real world. However, this type of testing does not demonstrate the complete vehicle system behavior in the same manner as in the real world. Complete vehicle system testing would provide the highest level of confidence that laboratory results accurately and fully reflect real-world performance and would be appropriate and suitable for regulators who do not have regular access to detailed technical information on specific types and models of vehicles. Furthermore, while vehicle emission, efficiency, and fuel consumption characteristics may degrade with the use of automated longitudinal control, systems can also be utilized to improve them. However, achieving these goals requires well-controlled laboratory-based test equipment and associated methods to maximize test accuracy and representativeness for both vehicle manufacturers and regulators.

[0010] Conventional laboratory tests for the purpose of emissions or energy efficiency compliance typically involve measuring the exhaust emissions or energy efficiency of a subject vehicle operating at one or more vehicle speed schedules on a dynamometer. The various vehicle speed schedules are intended to represent various types of real-world vehicle operation. For example, the Environmental Protection Agency (EPA) uses different speed schedules to represent urban operation, highway operation, and relatively aggressive vehicle operation. In each case, the vehicle is operated by a driver as close as possible to the corresponding speed schedule. However, relatively new vehicle models are increasingly using autonomous and dynamic longitudinal speed control as a convenient feature. These features are truly convenient because they automatically maintain a safe vehicle separation under any operating conditions ranging from congested urban traffic to highway operation, and because the cost of the technology is rapidly decreasing, it is likely that in the future, autonomous speed control features will be found and used in most vehicles. And they are likely to continue to be one of the key technologies for future fully autonomous vehicles. Summary of the Invention Problems to be Solved by the Invention

[0011] However, as the availability and use of autonomous vehicle speed control continue to grow, methods for testing such vehicles in laboratory settings, particularly in conventional test methods based on the behavior of subject vehicles in specific vehicle speed schedules using dynamometers, remain unclear. It is uncertain whether standard vehicle speed schedules will be used in the future for vehicles that control their own speed for most of the time. For example, a speed schedule may be simply interpreted as the speed at which the subject vehicle is compelled to follow in response to the speed of a leading vehicle moving at “traffic speed” in each case, or it may instead be interpreted as “traffic speed” itself (the traffic in which the subject vehicle is embedded and must follow). [Means for solving the problem]

[0012] overview In this specification, specific embodiments may relate to conducting laboratory tests of autonomous vehicles or vehicles with autonomous longitudinal speed or acceleration control to obtain accurate and repeatable exhaust gas mass emission measurements and energy efficiency measurements and automatic braking actions of any vehicle type in the case of ICE vehicles—measurements and actions representing real-world energy efficiency and tailpipe emissions—on any route and in any set of surrounding conditions of interest for any longitudinally controlled vehicle model. These embodiments provide apparatus and methods to allow for accurate determination of the impact on emissions and energy efficiency and automatic or emergency braking actions of autonomous longitudinally controlled vehicle functions. Furthermore, emissions, energy efficiency, and safety system performance can be calibrated, evaluated, and improved by simulating or replicating real-world traffic events in a controlled environment.

[0013] More specifically, certain embodiments relate to devices that simulate other vehicle traffic and roadside objects by generating appropriate vehicle databus signals representing the presence of traffic and objects, intercept and remove actual databus signals representing a view of objects based on vehicle sensors, and “inject” or replace the generated signals onto the databus in place of the original signals. The generated signals do not necessarily have to have a real-world basis or origin, but for some purposes it is advantageous that the generated signals are derived from real-world databus signals obtained by recording signals during previous road driving of the subject being simulated in the laboratory.

[0014] In the latter case, a vehicle with longitudinal speed control is operated along a desired route in the real world, with the vehicle speed and relative location or motion of objects detected by speed control sensors recorded from the vehicle data bus. The vehicle or powertrain is then moved into a laboratory or on a designated test route or track and operated with longitudinal speed control activated. If the test is conducted in a laboratory, the vehicle will be operated in conjunction with a dynamometer. In all cases, vehicle data bus signals including object detection or motion are removed and replaced with data bus signals that have been "converted" based on the difference between the real-world vehicle speed and the simulated vehicle speed history. The "converted" signals represent the relative location and motion that the same object would have been found to possess if the real-world speed was the same as the simulated speed or the speed during the test. As a result, vehicles with longitudinal speed control and powertrain calibration change during the test, which inevitably affects the vehicle's speed. This allows for retesting and direct comparison without the need to recreate the exact same traffic scenario and without returning to the real world after each test.

[0015] An apparatus is disclosed for transforming the relative location and motion of an object detected in advance based on a vehicle speed sensor or dynamometer feedback, intercepting and removing relevant data bus signals from a vehicle, and replacing the data bus signals with the transformed signals at that location.

[0016] Furthermore, methods are disclosed for converting the location or motion of a set of detected objects from one reference frame, such as the real world, to another reference frame, such as a laboratory dynamometer or a test track simulation. The objects may be static, dynamic, or a combination of both.

[0017] Also disclosed are test methods for testing vehicles having automated longitudinal speed control in relation to a laboratory dynamometer or a road or test track, and based on converted or simulated object location or motion based on speed feedback from the vehicle or dynamometer.

[0018] If emission data or energy consumption from a portable emission measurement system (PEMS) is optionally collected for its test method, including real-world operation, the PEMS data can be directly compared to corresponding emission or energy efficiency data collected during laboratory or other test simulations performed under identical conditions to ensure acceptable agreement. This optional "verification" process serves to demonstrate a high degree of confidence that both laboratory or simulation tests and real-world measurements are correct and reproducible. [Brief explanation of the drawing]

[0019] Brief explanation of the drawing [Figure 1] This is a rendering of the field of view schedule of relative real-world object detection locations, either recorded from the vehicle's previous real-world driving or generated as a simulated scenario for testing. [Figure 2]This shows a device for recording, storing, converting, and transmitting real-world driving data for vehicle testing. [Figure 3] This represents a vehicle with longitudinal speed control operating on a road in the real world. During operation, vehicle speed and object detection data bus signals from the longitudinal speed control sensors are recorded. Weather information is also recorded by an onboard weather station. [Figure 4] This shows a vehicle being tested in a laboratory setting under simulated vehicle load and weather conditions. It also shows a device that replaces the original signals with "converted" relative object location signals on the vehicle's data bus. [Figure 5] This is a field-of-view rendering comparing the transformed relative object locations and motions, which were substituted in laboratory tests, with the real-world relative object locations from which they were derived. [Figure 6] This illustrates the process of combining static object database location data in a simulated vehicle location with converted dynamic object or target database data, as observed from the same simulated vehicle location. [Modes for carrying out the invention]

[0020] Detailed explanation This specification describes various embodiments of the present invention. However, the disclosed embodiments are for illustrative purposes only, and other embodiments may have various alternative forms that are not expressly illustrated or described. The figures are not necessarily to scale, and some features may be exaggerated or minimized to show details of particular components. Accordingly, certain structural and functional details disclosed herein should not be construed as limitations, but only as representative grounds for teaching those skilled in the art to utilize the invention in various ways. Those skilled in the art will understand that various features illustrated and described with reference to any of the figures may be combined with features shown in one or more other figures to produce embodiments that are not expressly illustrated or described. The combinations of features illustrated provide representative embodiments for typical uses. However, for particular uses or implementations, various combinations and modifications of features that are consistent with the teachings of this disclosure may be desirable.

[0021] Figure 1 is a plot of the field of view (FOV) time schedule of the relative locations of a set of real-world objects detected by a longitudinal velocity control sensor and recorded from conventional real-world vehicle tests or generated as a simulated scenario for testing a vehicle in a laboratory. Time column 70 shows, for illustrative purposes, three consecutive test time steps t=t n 71, t=t n+1 72, t=t n+2 73, and the final time step t=t f It shows 74.

[0022] The base FOV column 76 shows the corresponding time step t=t n 71, t=t n+1 72, t=t n+2 73, and t=t fAn exemplary FOV plot showing sets of objects 78, 80, 82, 84 detected or so simulated by a vehicle longitudinal speed control system at 74 is shown. Objects within each FOV plot are numbered, for example, as n+2 for reference purposes, and the corresponding FOV plot 82 shows a set of six detected objects 86 located in front of the vehicle. n+2 Such as 73 etc., are numbered for reference, and the corresponding FOV plot 82 shows a set of six detected objects 86 located in front of the vehicle.

[0023] Furthermore, associated with each time step 71, 72, 73, 74 and their respective associated FOV plots 78, 80, 82, 84 are vehicle speed arrays 88, 90, 92, 94, which each have an overall vehicle speed history associated with the test time up to the test time under consideration. For example, the corresponding t = t n+2 FOV plot 82 having a set of six detected objects 86 located in front of the vehicle at time step t = t n+2 73, the figure shows an associated vehicle speed array [v n ...v n+2 92 having a complete set or schedule of vehicle speeds associated with these times.

[0024] For purposes of illustration, the t = t n FOV plot 78 shows four objects detected by the vehicle longitudinal speed control system, which are numbered 1 - 4 as shown in the figure. At t = t n+1 In subsequent sequential time steps showing FOV plots 80 and t = t n+2 FOV plot 82, five objects and six objects are detected respectively, and these are also numbered as 1 - 5 and 1 - 6 as shown in the figure. Since the time steps are sequential and assuming they are spaced in a temporally close state, due to the mostly static locations of objects 1 - 4 relative to each other and the vehicle, objects 1 - 4 are likely the same objects within all three FOVs, while objects 5 and 6 are likely, respectively, at time step t = tn+1 and t=t n+2 This is a newly detected object in t=t. f Objects 1 and 2 shown in FOV84 may be completely different objects if a large amount of time has passed in the test, especially if the vehicle has changed its lane.

[0025] The time step progression and associated FOV and vehicle speed arrays depicted in Figure 1 represent FOV and speed data that can be recorded from a test vehicle with effective longitudinal speed control, under real-world road driving conditions, and always following behind a "leading vehicle" that limits the speed of the "following" test vehicle. The driver control settings of the longitudinal speed control system may include speed setpoints and different following distance settings or calibration options, as commonly found in current vehicles. The longitudinal speed control system may also operate in relation to a "base calibration" or another pre-set calibration that the driver cannot normally change.

[0026] When a driver needs to change the following distance setting on a production vehicle, or when a vehicle manufacturer needs to change the longitudinal speed control system calibration or powertrain calibration during the vehicle development process, the result will likely be some change in the performance or responsiveness of the test vehicle in achieving and maintaining a setpoint speed, or in a manner in which the vehicle "actively" or accurately attempts to maintain a specific following distance as a function of vehicle speed behind any leading vehicle encountered during road driving. For example, a vehicle manufacturer may desire a relatively poor and active response to changes in the speed of the leading vehicle, or to changes in the road gradient experienced by the test vehicle, for the purpose of improved fuel efficiency (in the case of an internal combustion engine powertrain) or energy efficiency (in the case of a BEV). Therefore, it is advantageous for vehicle manufacturers to be able to evaluate "trial calibrations" for both the powertrain and longitudinal speed control systems by performing laboratory tests to determine the effect of the changed calibration on exhaust emissions and fuel efficiency in ICE-equipped vehicles and on the energy efficiency of BEVs.

[0027] In the case of a fixed set of real-world traffic conditions and roadside object locations over time, as described above, slightly different vehicle powertrain or longitudinal speed control system calibration (i.e., use of "trial calibration") is required for the first time step t n This will result in real-world driving time step recordings of the relevant FOV and vehicle speed arrays that differ when compared to those associated with the "base calibration" at all time steps after 71. Figure 1 illustrates an example of this behavior. Initial time step t n In this case, before any difference in motion shown by vehicles having different calibrations, as shown in the figure, the associated base calibration FOV 78 and trial calibration FOV 96 are identical, and the vehicle speed array [v n ]78 and [v' n ]100 are identical. However, they are different in subsequent time steps. For example, v n+1 >v'n+1 and, v n+1 >About v' n+1 If so, time step t n+1 The base FOV of 80 will likely show the same object as the trial-calibrated FOV of 97, but these will appear as if observed from a slightly more relative and advantageous point, as shown in the illustration. A similar reasoning explains the difference between the subsequent time-step FOVs shown.

[0028] Time-adjusted base calibration vehicle speed history [v n ...v f ]94 and trial calibration vehicle speed history[v' n ...v' f The calculation of the difference between 99 is permitted to be derived by a simple geometric transformation of the base calibrated FOVs 78, 80, 82, 84, on which the trial calibrated FOVs 96, 97, 98, 99 are recorded. For movement along the same one-dimensional path, any instance of time t n Velocity history array in [v1...v n ] and the associated actual location d and velocity history array [v'1...v' n The difference between the simulated location d' associated with ] is given by the following formula for small time steps: d-d'=[v1...v n ]dot{[t1...-t n ]-[t0...-t n-1 ]} This value represents small changes in the simulated vehicle's location relative to a real or simulated test vehicle (base vehicle), which is used, for example, to generate a database of relative object locations and motions during real-world driving with ACC system object detection logging or virtually generated simulated object locations and motions. The difference in the simulated vehicle's location relative to the base vehicle is why the object locations must be transformed over any given time step. The transformation process will be described in detail later.

[0029] Figure 2 shows an FOV simulation device 110 used to evaluate the impact of a vehicle's trial powertrain and longitudinal speed control system calibration on emissions and fuel consumption or energy efficiency, in comparison to a base calibration and in relation to a chassis, engine, or powertrain dynamometer for simulating real-world loads on the vehicle powertrain.

[0030] The FOV simulation device or FOV simulator 110 includes an enclosure 116, a processor 112, memory storage 114, a switch 152 for selecting the operating mode, a 12-volt power supply 144, and various external data cables for receiving data from the vehicle data bus, transmitting data to the vehicle data bus associated with the vehicle longitudinal speed control system, or transmitting and transferring data to and from a computer. The functions of the individual components will be described in more detail later.

[0031] The memory module 114 includes a target ID database memory 148 for storing vehicle longitudinal speed control system target object identification bus message identifiers and associated message structures; an FOV frame memory 146 for storing vehicle longitudinal speed control system sets based on the time or location at which the detected objects were acquired or simulated; and a vehicle speed history memory 150 for storing acquired or simulated vehicle speeds correlated with the FOV frames stored in the FOV frame memory 146.

[0032] The Vehicle Longitudinal Speed ​​Control Bus (VLSCB) cable 130 connects the FOV simulator 110 to the Vehicle Longitudinal Speed ​​Control System Bus, which provides detected object information, by using a VLSCB cable to the FOV simulator connector 132 at one end and a VLSCB cable to the object output data bus connector 136 at the other end. A VLSCB cable to the object input data bus connector 140 is located at the opposite end of the VLSCB cable. Vehicle speed data messages are transported from the vehicle data bus to the FOV simulator 110 via a speed data bus cable 118 connected from the vehicle, with a vehicle speed bus connector 122 and a vehicle speed FOV simulator connector 120 located at the opposite end.

[0033] When the vehicle under test is operating in conjunction with a dynamometer, the dynamometer controller speed signal cable 124 can optionally be used to transmit speed information from the dynamometer to the FOV simulator. Alternatively, the dynamometer controller speed signal cable 124 is connected to the dynamometer controller (not shown) using the dynamometer controller connector 128 and the FOV simulator dynamometer speed connector 126. As a result, the user can select the speed signal to use.

[0034] Power supply 144 provides 12VDC power to the FOV simulator 110 on power input line 142 so that it can function as a standalone device, and computer interface cable 160 is used to communicate with an external computer (not shown) by connecting the computer to computer interface connector 164 and FOV simulator computer connector 162. The following describes an illustrative usage of the FOV simulator device.

[0035] First, a vehicle equipped with longitudinal speed control 172 to be tested (e.g., BEV) or calibrated for emissions, fuel consumption, or energy efficiency is identified. The FOV simulator is connected to a host computer, and a complete set of vehicle-specific target object identification message identifiers and formats used by vehicle type 1 is uploaded to the FOV simulator 110 by processor 112 and stored in the target ID database 148. Vehicle-specific vehicle speed identification message identifiers and formats used by vehicle 1 are uploaded to the FOV simulator 110 in a similar manner and stored in the target ID database 148. Connection to an external computer is made using a computer interface cable 160 having an FOV simulator computer connector 162 at one end and a computer interface connector 164 at the other end. If it is necessary to use a virtual traffic or simulated driving scenario instead of a scenario recorded in the real world, the simulated FOV object detection frame and its corresponding vehicle speed and associated timestamp are uploaded to the FOV frame memory 146 and the vehicle speed history memory 150, respectively.

[0036] It may be advantageous to use real-world object detection scenarios and traffic conditions because, by definition, they are realistic and significant. The use case or "worst-case" scenario can be based on real-world behavior. In this case, switch 152 is set to the off or "0" position, causing the FOV simulator 110 to record all FOV object detection frames and vehicle speed frames from vehicle 1 as defined by the target ID database 148. The recorded FOV frame data and vehicle speed history data and associated timestamps are identified and stored by processor 112 in the FOV frame memory 146 and vehicle speed history memory 150, respectively.

[0037] Referring to Figure 3, Vehicle 1 can be driven in the real world on any target route while implementing any desired real-world traffic or object scenario based on the objectives of the test. Furthermore, optionally, vehicle emission and fuel consumption data or energy efficiency data can be acquired and stored, for example, by using PEMS4 in the case of emission and fuel consumption for ICE or hybrid electric vehicles, or by using a power consumption meter (not shown) in the case of energy efficiency for BEVs. Continuous measurement and recording of PEMS weather station data 30, including atmospheric pressure, ambient air temperature, and humidity, or the use of another weather station (not shown), are advantageous data to collect during operation to accurately replicate and simulate laboratory operation.

[0038] Referring again to Figure 2, during overall real-world driving, FOV object detection frame messages are identified, for example, by the processor 112 by comparing them with the target ID database 148 where input messages received on the VLSCB cable 130 are stored, and all of them are stored in the FOV frame memory 146. Similarly, vehicle speed messages, such as those received on the speed data bus cable 118, are identified by comparing them with the stored vehicle speed message format stored in the vehicle speed history memory 150 along with the relevant time. In relation to the above-described device, this can be achieved by setting switch 152 to the "0" position and connecting the vehicle longitudinal speed control bus input data line 170 to the vehicle longitudinal speed control bus output data line 156. When the object input data bus connector 140 is connected to vehicle 1 as described above, it carries all data messages from the object output data bus connector 136. Alternatively, the object input data bus connector 140 can remain disconnected from the vehicle 1 if the VLSCB cable 130 is not connected during installation.

[0039] As a result, by using data obtained from a separate computer or real-world driving, that is, by following either of the two options described above, the FOV simulator 110 will include a baseline set of data for real-world driving or the target traffic scenario in the memory storage 114, regardless of whether simulation scenarios are entered into the FOV frame memory 146 and the vehicle speed history memory 150.

[0040] If the FOV simulator 110 was used to record a baseline set of data during previous real-world driving, the FOV simulator 110 can be used later to supply and return the same data to vehicle 1 in relation to dynamometer load application and simulated atmospheric conditions based on real-world atmospheric conditions pre-recorded by the weather station. By doing so, the previous real-world control of the powertrain during real-world driving is replicated in the laboratory. Alternatively, the control of the powertrain during simulated driving or simulated driving scenarios or operations can be generated in the laboratory (or on the track).

[0041] Referring to Figures 2 and 4, the vehicle 1 or powertrain under test or calibration is coupled to the dynamometer by a normal test procedure, or, alternatively, may be tested on an isolated road or test track. In the case of a chassis dynamometer 10 with chassis roll 12, the appropriate load application using the dynamometer is achieved in a manner consistent with known dynamometer load application methods known in the art of engine, powertrain, and chassis dynamometer testing. If it is desirable to replicate real-world atmospheric conditions, the road load applied by the dynamometer can be adjusted, as known in the art, to compensate for differences in aerodynamic drag resulting from any differences in atmospheric pressure, temperature, and humidity between values ​​measured in the real world and the atmospheric conditions available for testing in the test cell 50. In the case of a vehicle with an ICE, an altitude / temperature / humidity simulator 57 can be used to ensure that the powertrain operates as it has operated or will operate in the real world. A capital-intensive alternative involves simulating the atmospheric conditions of the entire vehicle or powertrain by using a test cell 50 enclosed within an environmentally controlled chamber (not shown).

[0042] For example, a dynamometer speed signal, such as roll speed or an equivalent vehicle speed signal, is input from the dynamometer controller 11 to the FOV simulator 110 by connecting the FOV simulator 110 to the dynamometer controller using a speed signal cable 124. A vehicle data bus (not shown) is located thereto, carrying sensor-fused FOV objects identified by individual sensors or the vehicle longitudinal speed control system. The VLSCB cable 130 is connected to the FOV simulator 110 at one end and is connected to the vehicle longitudinal speed control system bus that provides detected object information by splicing or connecting the bidirectional VLSCB cable 130 into the bus by using the VLSCB cable to the object output database bus connector 136 on the data source side of the spliced ​​bus and the VLSCB cable to the object input data bus connector 140 on the data usage side of the spliced ​​bus. As a result, the FOV simulator 110 can intercept sensor-based FOV data generated in a static test environment, and the FOV simulator 110 can replace the converted FOV data based on the difference between the history of the dynamometer controller speed signal and the vehicle speed history memory 150, which is pre-loaded with real-world or simulated scenario data. A vehicle data bus carrying vehicle speed messages (not shown) is located and connected to the FOV simulator 110 using a speed data bus cable 118 with a vehicle speed bus connector 122, and the FOV simulator 110 is powered by connecting itself to a 12VDC voltage source from the vehicle or an independent power supply.

[0043] Before starting the dynamometer test, switch 152 is set to the "ON" or "1" position, causing the FOV simulator 110 to replace all FOV object detection frames input to the FOV simulator 110 from vehicle 1 with similar messages, as defined by the target ID database 148, in which case the relative location of the object is converted by processor 112 (Figure 2) based on the difference between the vehicle speed history and its correlated timestamp included in the set of speed data frame inputs on the speed data bus cable 118. The net result is that instead of the vehicle powertrain controller controlling the powertrain based on the vehicle longitudinal speed control system sensor FOV or fuse fused FOV and the actual ambient conditions in the test cell during the test, the powertrain controller control is configured to control the powertrain not only based on a pre-recorded or generated real-world FOV or simulated FOV, but also on a simulated ambient condition that simulates a desired ambient test condition that is equal to or different from a pre-recorded real-world ambient condition.

[0044] If the purpose of the laboratory or track test is to replicate conventional road driving or to evaluate calibration changes to the powertrain or longitudinal speed control system in comparison to the configuration used during conventional road driving, the driver control settings of the longitudinal speed control system are matched to real-world driving settings. Otherwise, other test values ​​are used depending on the purpose of the laboratory or track test. In the former case, the speed-controlled fan 25, the atmosphere simulator 57, or the weather-controlled chamber containing the test cell, as well as the dynamometer controller, are all set to a state corresponding to the start of the test in relation to time-stamped data in the FOV simulator memory module 114. To start the test sequence, a start switch (not shown) is pressed. Once pressed, FOV and vehicle speed data are acquired by the processor 112 from the memory module 114 in the order of their associated time stamps. All incoming FOV frames received by the FOV simulator 116 on the VLSCB cable 130 are identified by the processor 112 by referring to the target ID database memory 148. All other messages are immediately rebroadcast on the VLSCB output cable 134 by the FOV simulator.

[0045] The FOV frames are not simply rebroadcast literally. Sequentially and in a synchronized manner according to their timestamps, the FOV frames acquired from the memory module 114 are transformed to account for the difference in location (d-d') between the base vehicle and the simulated vehicle as time progresses. In this case as well, these location differences are due to the accumulated effect of all instantaneous vehicle speed differences up to the current test time. These are calculated by considering the time-dependent base speed history stored in the vehicle speed history memory 150 and the simulated speed history input on the speed signal cable 124 or the vehicle data bus cable 118, as shown above. More specifically, the relative geometric locations of detected objects included in the stored FOV frames are transformed to account for the difference in location along the test route. Replacement data frames are then broadcast on the FOV simulator VLSCB output cable 130 and to the vehicle 172, thereby replacing the original data frames from the vehicle 172 via the input cable 138 during the simulation.

[0046] The conversion of logged data from real-world or simulated driving of the aforementioned base vehicle is visually illustrated in Figure 5, and methodologically in more detail in Figure 6.

[0047] In a key case of an autonomous vehicle embedded in traffic control, namely, a case where the autonomous control system actively limits the speed of vehicle 1 due to the presence of other vehicles ahead and along its expected path, Figure 5 illustrates a general method for translating the locations of sets of objects and / or targets identified and recorded by the autonomous vehicle control system during real-world driving or generated for simulation. For illustrative purposes, the FOV from individual sensors of the autonomous vehicle control system, or, for example, a bird's-eye FOV, is shown when the vehicle is at location B215 in the figure. BLet us consider a set of objects and / or targets represented by a general set 207 of object symbols having output objects in a given time instance from a sensor-fused bird's-eye FOV such as 209. For the sake of drawing simplicity, the FOV B Let us assume that 209 has, for example, an inner moving target FOV 201 indicating the location and speed of other nearby vehicles, and an outer object FOV 217 indicating the location of static objects, including, for example, road signs and guardrails, without limitation. In this simplified case, it is easy to graphically illustrate the overall transformation method, as shown in the figure.

[0048] Figure 5 depicts two illustrative vehicle locations, location A213 and location B215, which are traversed in previous real-world driving or virtual or simulated driving. In previous real-world or simulated road driving, vehicle 1 first uses a bird's-eye FOV B Bird's-eye view before arriving at location B215, associated with 209 A It was located at location A213 (the whole location is not shown). The parameters recorded or simulated in the real world associated with each location are defined below. Location A213: Bird's-eye field of view A (The whole thing is not shown) is the static object FOV sj 225 and Mobile Object FOV mj (Not shown in the diagram), that is, it has the relative location and velocity of the combined static and dynamic objects, t rj 221 is the time when vehicle 1 arrives at location A213. d rj 223 is the cumulative distance driven to location A213, and, φ rj 229 is the vehicle heading at location A213. Location B215: Bird's-eye field of view B 209 is static FOVsj 217 and Mobile Object FOV mj 201, that is, the relative location and velocity of the combined static and dynamic objects, t ri 233 is the time when vehicle 1 arrives at location B215. d ri 235 is the cumulative distance driven to location B215, and, φ ri 231 is the vehicle heading at location B215. The following variable is defined as AvB, which is associated with the difference between location A213 and location B215. AvB: The heading difference θ241 is given by θ = φri - φrj, The straight line between location A213 and location B215 is defined by r243, and, The actual route between location A213 and location B215 is defined as ΔS247.

[0049] Once the real-world driving or real-world simulation is complete, the parameters recorded in the real world or simulation as described above include a database 301 (Figure 6) of all objects and targets, and their speed or motion detected by the vehicle 1 autonomous control system for the entire real-world driving or simulation. Whether each object in the database 301 is static or dynamic, while it is detected with reference to a local reference frame, each object is actually similarly positioned in relation to a global coordinate system which can be readily calculated by those skilled in the art. Regardless of whether the objects and their motion are observed from the location and heading of vehicle 1 when they are detected, or from a different or hypothetical location of vehicle 1, changes in their location and time must be identical within any global reference frame. This fact can be used to the advantage of testing any vehicle with an autonomous or adaptive cruise control system, in particular to evaluate changes to the calibration of the autonomous control system or changes to the powertrain of the same vehicle 1 used to generate the database 301, or other minor changes made to the system. The reason behind this is that, for example, small changes to the overall performance of vehicle 1, such as calibration changes, usually result in relatively small changes in performance and, in cases where vehicle 1 is always controlled by the same combination of leading vehicle location and speed and / or driver-selectable maximum speed settings, small changes in the vehicle's global location over time in a balanced manner.

[0050] The location and velocity or motion of a set of objects to be recorded, detected or observed by the autonomous system of Vehicle 1 during real-world or simulated road driving and with reference to a local reference frame, can be transformed by those skilled in the art from different global locations and coordinate system orientations to different views of the same set of objects. For example, the location and motion of objects to be recorded, detected by the autonomous control system of Vehicle 1 at location B215 while operating under a first autonomous control system or powertrain calibration, can be easily transformed into a view that the same autonomous control system of Vehicle 1 would have if operating under a second autonomous control system or powertrain calibration different from the first, which would consequently result in the location of Vehicle 1 becoming location A at the same route and at the same test time due to performance differences.

[0051] The above transformation can be symbolically described as follows:

number

[0052] Figure 6 shows a specific example of the conversion process described above performed by the processor 112 of the FOV simulator 110 while a vehicle 1 equipped with autonomous or adaptive cruise control is operating as described above, with trials or test calibrations, and in relation to a laboratory chassis, powertrain, or engine dynamometer. Database 301 contains all autonomous vehicle control system sensor data or fused sensor data collected during real-world driving by the FOV simulator 110, or by any other device used to collect similar data or data generated for simulated driving as described above, while switch 152 is in “record” mode, i.e., position “0”. With the FOV simulator 110 switched to position “1”, the conversion is performed based on the difference between the actual location of the vehicle, which is represented in one dimension by the dynamometer distance d303 driven based on the dynamometer or vehicle speed feedback as described above, and in the following manner:

[0053] In the case of a trial calibration test case in any dynamometer distance d303 driven of 1.19 miles, the corresponding set of static object locations S AS 203 is selected. Vehicle 1 reaches this location with base calibration at t=4 seconds 205, while Vehicle 1 reaches this same location with trial calibration at t'=5 seconds. Referring to database 301, the location corresponding to the same base calibration test time t, i.e., t=t'=5 seconds, is S Bd The dynamic object identified as 207 has a corresponding set of location and velocity, with d = 1.23 miles. Then, S Bd 207 is transformed by the above-described method and S is used to form a complete set of objects and targets for presentation to the autonomous vehicle control system using the above-described data intercept method. ASIn combination with 203, that is, FOV data based on sensors generated in a static laboratory test environment is intercepted by the FOV simulator 110 and replaced by the converted FOV data by the FOV simulator 110.

[0054] Those skilled in the art know that driving style affects the energy efficiency of all types of automotive powertrains and the exhaust emission levels of powertrains utilizing internal combustion engines (ICE). Accelerator pedal motion, braking, and powertrain calibration all affect vehicle efficiency and emissions. However, autonomous longitudinal speed control features on vehicles, such as adaptive cruise control (ACC), replace the driver's normal accelerator pedal input and braking activities, which may rely on different sets of powertrain calibrations while the vehicle is operating autonomously.

[0055] The test laboratory may be housed in a weather-controlled chamber (not shown) in which ambient air conditions of pressure, temperature, and humidity can be individually controlled. Such a laboratory allows for the replication or simulation of ambient air conditions to reflect previous real-world operating conditions to be replicated or simulated in the laboratory, or simulation of specific ambient air conditions of a subject during tests, intended to be identical except for the controlled variables, for the purpose of maximizing the accuracy of the test results.

[0056] Alternatively, a relatively cost-effective implementation for replicating or simulating dynamic ambient air conditions utilizes a “conditional environment simulator” 52, which has recently become commercially available. This provides a less capital-intensive means of dynamically changing the ambient conditions experienced by the powertrain during testing and replicating and simulating desirable ambient air conditions, while allowing the use or continuous use of a standard emission test laboratory. In this case, ambient air pressure, temperature, and humidity conditions are generated by the conditional environment simulator 52 and applied only to the powertrain and necessary vehicle sensors by connecting the conditional environment simulator 52 to the subject vehicle 6 engine intake system (not shown) via an intake hose 56 and to the vehicle's tailpipe 24 via an exhaust gas hose 54. The conditional environment simulator 52 controls the intake pressure, exhaust back pressure, and intake humidity to appropriately fixed selected values ​​or programmatically controlled dynamic values, or to mimic conditions recorded during real-world testing in a state appropriately synchronized with the speed of the subject vehicle 6 on the dynamometer assembly 12.

[0057] Any of the illustrated example test methods may, at their discretion, utilize an ambient environment simulator, or alternatively, a full-vehicle "environmental test chamber," to conduct accurate tests when the desired weather conditions differ from the ambient conditions of the available laboratory.

[0058] Before operating the vehicle on the dynamometer, appropriate road load parameters for the test vehicle, appropriate road gradient parameters for the route, and dynamic environmental conditions are selected and programmed into the environmental control system, i.e., into the control system for the relevant environmental chamber or powertrain "environmental condition simulator." For example, in cases where the laboratory environmental conditions differ significantly from the desired environmental conditions, such as when simulating real-world road driving to control traffic in different weather conditions, it is important to ensure that the powertrain operates within an appropriate calibration space and thereby maintain appropriate environmental conditions to produce representative emissions and demonstrate representative energy efficiency.

[0059] The converted data envisioned above can be used by a vehicle to perform various control actions such as braking, steering, and acceleration in order to navigate a road (real or simulated) while simultaneously avoiding contact with other objects (real or simulated). For example, if the converted data indicates an approaching object (real or simulated), the vehicle can respond with appropriate braking and / or steering actions.

[0060] The processes, methods, or algorithms disclosed herein may be supplied to and implemented by processing units, controllers, or computers, which may include existing programmable electronic control units or dedicated electronic control units. Similarly, the processes, methods, or algorithms may be stored as data and instructions executable by controllers or computers in many forms, including, but not limited to, information permanently stored on a non-writable storage medium such as a read-only memory (ROM) device and information modifiable on a writable storage medium such as a floppy disk, magnetic tape, compact disk (CD), random access memory (RAM) device, and other magnetic and optical media. Alternatively, the processes, methods, or algorithms may be implemented in a software executable object. Or, instead, the processes, methods, or algorithms may be implemented holistically or in part using appropriate hardware components such as application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), state machines, controllers, or other hardware components or devices, or combinations of hardware, software, and firmware components.

[0061] The terms used herein are illustrative, not limiting, and should be understood to be subject to various modifications without exceeding the spirit and scope of this disclosure.

[0062] As described above, features of various embodiments can be combined to form further embodiments, which may not be explicitly described or illustrated. While various embodiments may be described as offering advantages or being preferred over other embodiments or prior art implementations in relation to one or more desirable characteristics, those skilled in the art will recognize that one or more features or characteristics may be compromised in achieving desirable overall system attributes that depend on a particular application and implementation. These attributes include, without limitation, cost, strength, durability, life cycle cost, marketability, appearance, packaging, size, serviceability, weight, manufacturability, ease of assembly, etc. Therefore, embodiments described as undesirable over other embodiments or prior art implementations in relation to one or more characteristics are also included in the scope of this disclosure and may be desirable for a particular application.

Claims

1. A vehicle testing device, It has a processor, The aforementioned processor, (i) Location or motion data derived from the vehicle sensor that defines the location or motion of one or more objects detectable by the vehicle sensor, and in response to location or motion data defined in relation to a first local reference frame associated with the vehicle at a first vehicle location, generate a converted set of location or motion data that defines the location or motion of the one or more objects in relation to a second local reference frame associated with the vehicle at a second vehicle location different from the first vehicle location, based on at least one of the test vehicle signal, dynamometer speed feedback signal, steering angle signal, or GPS heading signal, such that the location or motion data and the converted location or motion data represent the same location or motion of the one or more objects in a common global reference frame, and (ii) Outputting the converted set of location or motion data to the vehicle in order to have the vehicle perform a control operation or to evaluate the operation of the vehicle control system during the test, A vehicle testing device programmed to do so.

2. The vehicle test apparatus according to claim 1, wherein the processor is further programmed to derive the set of location or motion data from the outputs of one or more vehicle sensors and to record the corresponding location or motion of the test vehicle in the global reference frame based on a vehicle speed signal or GPS signal.

Images