Method and apparatus for longitudinal control of autonomous vehicle

Abstract

The present disclosure provides a method for longitudinal control of an autonomous vehicle, comprising: determining, based on driving environment perception data, that at least one potential cut-in vehicle is present in front of the autonomous vehicle; obtaining a cut-in probability of each potential cut-in vehicle in the at least one potential cut-in vehicle; selecting at least one target vehicle based on the cut-in probability; and calculating an expected following distance for the at least one target vehicle.

Description

Technical Field

[0001] The present disclosure generally relates to the field of autonomous vehicles, and more particularly to methods and devices for longitudinal control of autonomous vehicles. Background Art

[0002] Currently, vehicle autonomous driving technology is developing rapidly. Longitudinal control is one of the core technologies of autonomous driving. Longitudinal control refers to the control of an autonomous vehicle in the driving direction, including, for example, controlling the vehicle speed of the autonomous vehicle and / or controlling the distance between the autonomous vehicle and the vehicle in front by coordinating the drive system and the braking system of the autonomous vehicle. Some autonomous vehicles use Adaptive Cruise Control (ACC) technology to achieve longitudinal control. Using ACC technology, when there is no vehicle in front in the current lane of the autonomous vehicle or the vehicle in front is far away, the autonomous vehicle can travel at a set speed. Generally, such an ACC control mode can be referred to as a constant speed cruise mode. When there is a vehicle in front in the current lane or the vehicle in front is close, the vehicle in front can be selected as a following target, and the vehicle speed of the autonomous vehicle can be automatically controlled based on the driving conditions of the following target so that the autonomous vehicle always maintains a safe distance from the following target. Generally, such an ACC control mode can be referred to as a following mode. Using ACC technology, by adaptively switching between the constant speed cruise mode and the following mode based on road conditions, stable and safe assisted driving can be ensured while reducing the workload of the driver.

[0003] However, conventional ACC technology only selects a single target to perform longitudinal control. When the selected target suddenly switches, for example, when a vehicle in front in an adjacent lane suddenly cuts into the current lane or other similar situations occur, the autonomous vehicle may need to perform emergency braking to ensure a safe distance from the switched target, which may result in an unsatisfactory riding experience and even may lead to safety problems. Summary of the Invention

[0004] The present disclosure provides an improved mechanism for longitudinal control of an autonomous vehicle. The radar control unit of the autonomous vehicle can simultaneously monitor the driving conditions of multiple preceding vehicles to determine a target vehicle and calculate the desired following distance between the autonomous vehicle and the target vehicle. Then, the radar control unit can provide the determined target vehicle and the calculated desired following distance to the vehicle control unit (VCU). The VCU can then cause the autonomous vehicle to maintain the calculated desired following distance from the target vehicle by controlling the acceleration and deceleration of the autonomous vehicle. This improved longitudinal control mechanism avoids the problem of emergency braking caused by sudden switching of the target, which affects the riding experience when only focusing on one target.

[0005] According to one aspect of the present disclosure, there is provided a method for longitudinal control of an autonomous vehicle, including: determining, based on driving environment perception data, that there is at least one potential cutting-in vehicle in front of the autonomous vehicle; obtaining the cutting-in probability of each potential cutting-in vehicle among the at least one potential cutting-in vehicle; selecting at least one target vehicle based on the cutting-in probability; and calculating a desired following distance for the at least one target vehicle.

[0006] According to another aspect of the present disclosure, there is provided an apparatus for longitudinal control of an autonomous vehicle, including: a memory and a processor. The processor is coupled to the memory and is configured to execute the method according to any one of the various embodiments of the present disclosure.

[0007] According to still another aspect of the present disclosure, there is provided a computer-readable medium storing a computer program including instructions that, when executed by a processor, cause the processor to be configured to execute the method according to any one of the various embodiments of the present disclosure.

[0008] According to another aspect of the present disclosure, there is provided a computer program product including instructions that, when executed by a processor of a computing device, cause the processor to be configured to execute the method according to any one of the various embodiments of the present disclosure. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] Various embodiments of the claimed subject matter will now be described by way of example with reference to the accompanying drawings. In the different drawings, the same reference numerals are used to denote the same or similar components.

[0010] Figure 1 FIG. shows a schematic diagram of the overall architecture of a longitudinal control mechanism for an autonomous vehicle according to an exemplary embodiment of the present disclosure.

[0011] Figure 2A schematic diagram showing an example scenario without a current following vehicle according to an example embodiment of the present disclosure.

[0012] Figure 3 A schematic diagram showing an example scenario with a current following vehicle according to an example embodiment of the present disclosure.

[0013] Figure 4 A flowchart showing a method for longitudinal control of an autonomous vehicle according to an example embodiment of the present disclosure.

[0014] Figure 5 A block diagram showing a device that can implement a method for longitudinal control of an autonomous vehicle according to an example embodiment of the present disclosure. Detailed implementation

[0015] In the following description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure. However, those skilled in the relevant art will recognize that the present disclosure may be practiced without one or more of the specific details, or may be practiced using alternative methods, components, etc. In some instances, well-known structures, operations are not shown or described in detail so as not to unnecessarily obscure the present disclosure.

[0016] Figure 1 A schematic diagram showing the overall architecture 100 of a longitudinal control mechanism for an autonomous vehicle according to an example embodiment of the present disclosure.

[0017] As Figure 1 shown, the radar control unit 102 of the autonomous vehicle can receive driving environment perception data 104 to determine whether there is at least one potential cut-in vehicle in front of the autonomous vehicle. A potential cut-in vehicle refers to a vehicle in an adjacent lane to the current lane where the autonomous vehicle is located and may cut into the current lane to become a following target.

[0018] The driving environment perception data 104 can include various sensor data from driving environment perception sensors. The driving environment perception sensors can include radar, lidar, cameras, and any other form of sensors that can generate data associated with the current driving environment of the autonomous vehicle. The driving environment perception data 104 can characterize the conditions of each lane of the road, the driving conditions of the autonomous vehicle, and the driving conditions of other nearby vehicles.

[0019] The radar control unit 102 can also obtain the cut-in probability of each potential cut-in vehicle. The cut-in probability indicates the likelihood of a potential cut-in vehicle cutting into the current lane.

[0020] In one example, a probability prediction model can be utilized to generate the cut-in probability of each potential cut-in vehicle. The probability prediction model can be an artificial intelligence (AI) model for performing a regression task, which can predict the cut-in probability of each potential cut-in vehicle based on driving environment perception data. In one example, the driving environment perception data input into the probability prediction model can include data associated with the following: lane lines, the lateral distance between the potential cut-in vehicle and the autonomous vehicle, and the lateral acceleration between the potential cut-in vehicle and the autonomous vehicle. Existing open-source data (e.g., Apollo data) can be used to train the probability prediction model so that the trained probability prediction model can generate accurate cut-in probabilities for each potential cut-in vehicle.

[0021] In one example, the probability prediction model can be deployed on the radar control unit 102. In this case, the radar control unit 102 can utilize the probability prediction model to perform a probability prediction process to obtain the cut-in probability of each potential cut-in vehicle. In another example, the probability prediction model can also be deployed on any other computing device (e.g., the vehicle control unit, the roadside computing unit, etc.) to take advantage of its more powerful computing capabilities. In this case, the radar control unit 102 can receive the cut-in probability of each potential cut-in vehicle from the above-mentioned other computing device.

[0022] The radar control unit 102 can also obtain a maximum probability threshold and a minimum probability threshold associated with the cut-in probability. The maximum probability threshold and the minimum probability threshold can be used by the radar control unit 102 to determine the desired following distance. In one example, the maximum probability threshold and the minimum probability threshold can be determined based on existing open-source data (e.g., Apollo data). In one example, a self-learning algorithm can be used to determine the maximum probability threshold and the minimum probability threshold. In one example, a technician can manually set the maximum probability threshold and the minimum probability threshold.

[0023] The radar control unit 102 can select at least one target vehicle based on the cut-in probability, and then can calculate the desired following distance for the selected at least one target vehicle. Details regarding selecting the target vehicle and calculating the desired following distance will be further discussed in detail below in conjunction with Figure 2 and Figure 3 the two example scenarios shown.

[0024] The radar control unit 102 and the VCU 108 can be communicatively coupled to the CAN bus unit 106 respectively. The radar control unit 102 can provide information related to at least one selected target vehicle and the calculated desired following distance to the VCU 108 through the VCU 108. The VCU 108 can generate a longitudinal control command based on the received information related to at least one selected target vehicle and the calculated desired following distance to longitudinally control the autonomous vehicle. The longitudinal control command can include controls for the speed, acceleration, etc. of the autonomous vehicle. The VCU 108 can transmit the generated longitudinal control command to the power and brake control unit 110. The power and brake control unit 110 can execute the longitudinal control command, for example, accelerating or decelerating the autonomous vehicle. In one example, the power and brake control unit 110 can include a power system, an ESP system, an ABS system, etc.

[0025] Figure 2 and Figure 3 Two example scenarios are shown to facilitate the understanding of the details regarding selecting a target vehicle and calculating a desired following distance. In Figure 2 the shown example scenario, there is no current following vehicle in the current lane where the autonomous vehicle is located. This example scenario can correspond to the ACC cruise control mode. In Figure 3 the shown example scenario, there is a current following vehicle in the current lane where the autonomous vehicle is located. This example scenario can correspond to the ACC following mode. The radar control unit can determine whether there is such a current following vehicle based on the driving environment perception data. In the case of determining that there is no current following vehicle, the radar control unit selects to use the mechanism discussed in conjunction with Figure 2 to select a target vehicle and calculate a desired following distance. In the case of determining that there is a current following vehicle, the radar control unit selects to use the mechanism discussed in conjunction with Figure 3 to select a target vehicle and calculate a desired following distance.

[0026] Figure 2 A schematic diagram of an example scenario 200 without a current following vehicle according to an example embodiment of the present disclosure is shown.

[0027] Figure 2 An autonomous vehicle 202 is shown, which is in the current lane 204. The radar control unit can determine the adjacent lanes of the current lane 204 based on the driving environment perception data, for example, the left adjacent lane 206 and the right adjacent lane 208. The radar control unit can also determine, based on the driving environment perception data, whether there is a potential cut-in vehicle in front of the autonomous vehicle 202. A potential cut-in vehicle refers to a vehicle located in the adjacent lane 206 or 208 and in front of the autonomous vehicle, and thus may cut into the current lane 204.Figure 2 Exemplarily shown are two potential cut-in vehicles respectively located on the left adjacent lane 206 and the right adjacent lane 208.

[0028] The radar control unit can obtain the cut-in probability of each potential cut-in vehicle. As combined above with Figure 1 discussed, the cut-in probability of each potential cut-in vehicle can be generated by a probability prediction model based on driving environment perception data. Then, the radar control unit can select the vehicle with the maximum cut-in probability among the potential cut-in vehicles as the target vehicle.

[0029] As Figure 2 exemplarily shown, it can be determined that, for example, the cut-in probability of the potential cut-in vehicle 210 on the left adjacent lane 206 is the maximum, so this potential cut-in vehicle 210 can be selected as the target vehicle. It should be understood that Figure 1 the scenario where there is one potential cut-in vehicle respectively on the shown left adjacent lane 206 and right adjacent lane 208 is only an example. In other examples, there may be more or fewer potential cut-in vehicles on each lane, or potential cut-in vehicles at other positions, without affecting the implementation of the above target vehicle determination mechanism. It should be understood that in any example, the radar control unit can similarly select the potential cut-in vehicle with the maximum cut-in probability as the target vehicle based on the cut-in probability of each potential cut-in vehicle.

[0030] After selecting the target vehicle (e.g., 210), the radar calculation unit can calculate the desired following distance d desire . In Figure 2 the example scenario shown where there is no current following vehicle, the desired following distance d desire represents the desired distance between the autonomous vehicle 202 and the selected target vehicle 210.

[0031] The cut-in probability p of the target vehicle 210 can be compared with the maximum probability threshold and / or the minimum probability threshold combined above with Figure 1 discussed, and the desired following distance d desire is calculated based on the comparison result.

[0032] In one example, if the cut-in probability p of the target vehicle 210 falls within the threshold interval (p min , p max ) defined by the minimum probability threshold and the maximum probability threshold, that is, p min < p < p max , then the normalized cut-in probability pp of the target vehicle 210 and the single-target following distance d single for the target vehicle 210 can be calculated. The normalized cut-in probability pp can be obtained by normalizing the cut-in probability p within the range defined by the minimum probability threshold pmin and the maximum probability threshold p max Linear interpolation is performed within the defined threshold interval to calculate the normalized cut-in probability pp. Through the above linear interpolation operation, the normalized cut-in probability pp can be mapped to the data range (0,1). Single target following distance d single Refers to the expected following distance from the single target calculated when the target vehicle is regarded as a single target in the conventional longitudinal path planning mechanism. In one example, the single target following distance d can be determined based on the time headway to the target vehicle 210 and the current driving speed of the autonomous driving vehicle 202. single The time headway for the target vehicle 210 may be a fixed time headway or a headway headway between the target vehicle 210 and the autonomous driving vehicle 202. Then, the target vehicle 210 may be determined based on the normalized cut-in probability pp of the target vehicle 210 and the single target following distance d of the target vehicle 210. single To calculate the expected following distance d desire In one example, the desired following distance d may be desire = ... single The product of, that is:

[0033] d desire =pp*d single

[0034] The above mechanism realizes dynamic adjustment of the expected following distance based on the cut-in probability. For example, when the cut-in probability of the target vehicle is small, the calculated expected following distance is also small, which is beneficial to reduce the occurrence of emergency braking.

[0035] In one example, as the autonomous driving vehicle 202 gradually approaches the target vehicle 210, the cut-in probability p of the target vehicle 210 gradually increases and is greater than the maximum probability threshold p max (corresponding to the case where the target vehicle 210 cuts into the current lane 204), then the single target following distance d for the target vehicle 210 can be used. single To calculate the expected following distance d desire ,in:

[0036] d desire =d single

[0037] In one example, as the autonomous driving vehicle 202 gradually approaches the target vehicle 210, the cut-in probability p of the target vehicle 210 gradually decreases and is less than the minimum probability threshold p min (corresponding to the situation where the autonomous driving vehicle overtakes the target vehicle 210), the target vehicle 210 is no longer considered, and the vehicle can continue to travel in the cruise control mode.

[0038] Figure 3 FIG. 300 is a schematic diagram showing an example scenario in which there is a current following vehicle according to an example embodiment of the present disclosure.

[0039] Figure 3 An autonomous vehicle 302 is shown, which is on the current lane 304. There is also a current following vehicle 310 on the current lane 304, wherein the autonomous vehicle 302 is traveling in a following mode following the current following vehicle 310. The radar control unit can determine the adjacent lanes of the current lane 304 based on the driving environment perception data, for example, the left adjacent lane 306 and the right adjacent lane 308. The radar control unit can also determine whether there are potential cut-in vehicles on the left adjacent lane 306 and the right adjacent lane 308 respectively based on the driving environment perception data. A potential cut-in vehicle refers to a vehicle located on an adjacent lane and between the autonomous vehicle 302 and the current following vehicle 310, and thus may cut into the current lane 304. Figure 3 Exemplarily shown therein are a first potential cut-in vehicle 312 on the left adjacent lane 306 and a second potential cut-in vehicle 314 on the right adjacent lane 308. In one example, if there are multiple potential cut-in vehicles on any adjacent lane, only the potential cut-in vehicle with the highest cut-in probability may be considered. As Figure 3 shown, the first potential cut-in vehicle 312 is located between the autonomous vehicle 302 and the current following vehicle 310, and the second potential cut-in vehicle 314 is located between the autonomous vehicle 302 and the first potential cut-in vehicle 312.

[0040] The radar control unit can select the current following vehicle 310, the first potential cut-in vehicle 312, and the second potential cut-in vehicle 314 as target vehicles. Then, the radar calculation unit can calculate the desired following distance d for the target vehicles desire . In Figure 3 the example scenario where there is a current following vehicle 310 shown, the desired following distance d desire represents the desired distance between the autonomous vehicle 302 and the current following vehicle 310.

[0041] Then, the cut-in probability p1 of the first potential cut-in vehicle 312 and the cut-in probability p2 of the second potential cut-in vehicle 314 can be compared with the maximum probability threshold p Figure 2 and / or the minimum probability threshold p max and / or the minimum probability threshold p min in a similar manner as discussed in connection with desire .

[0042] In one example, if the cut-in probability p1 of the first potential cut-in vehicle 312 and the cut-in probability p2 of the second potential cut-in vehicle 314 both fall within the threshold interval defined by the minimum probability threshold p min and the maximum probability threshold p max That is, p min < p1 < p max and p min < p2 < p max , then the expected following distance d single-0 can be calculated based on the single-target following distance d desire for the current following vehicle 310, the normalized cut-in probability pp1 of the first potential cut-in vehicle 312, the normalized cut-in probability pp2 of the second potential cut-in vehicle 314, the first incremental distance Δd1 for the first potential cut-in vehicle 312, and the second incremental distance Δd2 for the second potential cut-in vehicle. desire In one example, the expected following distance d

[0043] can be calculated based on the following formula: desire d single-0 = d

[0044] + pp1 * Δd1 + pp2 * Δd2 single-0 The single-target following distance d Figure 2 for the current following vehicle 310 can be calculated in a similar way to the calculation of d single discussed above.

[0045] In one example, to calculate the normalized cut-in probability pp1 of the first potential cut-in vehicle 312 and the normalized cut-in probability pp2 of the second potential cut-in vehicle 314, the cut-in probability p1 of the first potential cut-in vehicle 312 and the cut-in probability p2 of the second potential cut-in vehicle 314 can be determined respectively through the probability prediction model described above, and the probability p0 of continuing to follow the current following vehicle 310 can be determined based on the following formula:

[0046] p0 = min(1 - p1, 1 - p2)

[0047] In this way, if any potential cut-in vehicle has cut into the current lane 304, so that its cut-in probability is equal to 1, then p0 can be directly determined to be 0 through simple calculation, in order to save computing resources and improve operation efficiency. In another example, the probability p0 of continuing to follow the current following vehicle 310 can also be determined through the probability prediction model described above. Then, for the cut-in probabilities p0, p1, p2 within the range defined by the minimum probability threshold p min and the maximum probability threshold p maxLinear interpolation is performed within the defined threshold interval to calculate the normalized cut-in probability pp0 of the current following vehicle 310, the normalized cut-in probability pp1 of the first potential cut-in vehicle 312, and the normalized cut-in probability pp2 of the second potential cut-in vehicle 314. Through the above linear interpolation operation, the normalized cut-in probabilities pp0, pp1, and pp2 can be mapped to the data range (0, 1) respectively, and pp0+pp1+pp2=1 is satisfied.

[0048] The first incremental distance Δd1 and the second incremental distance Δd2 can be calculated as follows:

[0049] Δd1=d single-1 -d1+d0-d single-0

[0050] Δd2=d single-2 –d2+d0-d single-0

[0051] where d single-1 is the single target following distance for the first potential cutting-in vehicle 312 and can be combined with the above Figure 2 Discussion single The calculation method is similar to that of single-2 is the single target following distance for the second potential cutting-in vehicle 314 and can be combined with the above Figure 2 Discussion single The actual distance between the autonomous driving vehicle 302 and the first potential cutting-in vehicle 312 is calculated in a similar manner, d1 is the actual distance between the autonomous driving vehicle 302 and the second potential cutting-in vehicle 314, d2 is the actual distance between the autonomous driving vehicle 302 and the second potential cutting-in vehicle 314, and d0 is the actual distance between the autonomous driving vehicle 302 and the current following vehicle 310.

[0052] In one example, as the autonomous driving vehicle 302 gradually moves forward, the cut-in probability p1 of the first potential cut-in vehicle 312 gradually increases and is greater than the maximum probability threshold p max , that is, p1>p max (corresponding to the case where the first potential cutting-in vehicle 312 cuts into the current lane 304), then based on the single target following distance d for the first potential cutting-in vehicle 312, the following formula is: single-1 , the normalized cut-in probability pp2 of the second potential cut-in vehicle 314 , and the second incremental distance Δd2 of the second potential cut-in vehicle 314 to calculate the expected following distance d desire :

[0053] d desire =d single-1 +pp2*Δd2

[0054] In one example, if as the autonomous vehicle 302 gradually moves forward, the cut-in probability p2 of the second potential cut-in vehicle 314 gradually increases and is greater than the maximum probability threshold p max , that is, p2 > p max (corresponding to the situation where the second potential cut-in vehicle 314 cuts into the current lane 304), then based on the single-target following distance d for the second potential cut-in vehicle 314 according to the following formula single-2 to calculate the expected following distance d desire :

[0055] d desire = d single-2

[0056] In one example, if as the autonomous vehicle 302 gradually moves forward, the cut-in probabilities p1 of the first potential cut-in vehicle 312 and p2 of the second potential cut-in vehicle 314 gradually decrease and are both less than the minimum probability threshold p min (corresponding to the situation where the autonomous vehicle overtakes the first potential cut-in vehicle 312 and the second potential cut-in vehicle 314), then the first potential cut-in vehicle 312 and the second potential cut-in vehicle 314 are no longer considered, but continue to follow the current following vehicle 310, and based on the single-target following distance for the current following vehicle 310 according to the following formula to calculate the expected following distance d desire :

[0057] d desire = d single-0

[0058] Although the above discussion in combination with Figure 3 describes an example scenario where there are two potential cut-in vehicles, when there is only one potential cut-in vehicle, the above discussion still applies. For example, the discussion about the first potential cut-in vehicle or the discussion about the second potential cut-in vehicle can be considered alone.

[0059] Figure 4 FIG. shows a flowchart of a method 400 for longitudinal control of an autonomous vehicle according to an example embodiment of the present disclosure. The steps of method 400 can be executed by the radar control unit discussed above in combination with Figure 1 discussion.

[0060] In step S402, based on the driving environment perception data, it can be determined that there is at least one potential cut-in vehicle in front of the autonomous vehicle.

[0061] In step S404, the cut-in probability of each potential cut-in vehicle among the at least one potential cut-in vehicle can be obtained.

[0062] In step S406, based on the cut-in probability, at least one target vehicle can be selected.

[0063] At step S408, an expected following distance can be calculated for the at least one target vehicle.

[0064] Figure 5 FIG. shows a block diagram of a device 500 according to an example embodiment of the present disclosure, which can implement a method for longitudinal control of an autonomous vehicle.

[0065] The exemplary device 500 includes a processor 504 connected to an internal communication bus 502. The processor 504 is configured to execute instructions in a memory 506 to implement the method for longitudinal control of an autonomous vehicle described in detail above. Examples of the processor 504 may include a central processing unit (CPU), a microcontroller, and the like. The memory 506 adapted to tangibly embody computer program instructions and data includes various forms of memories, such as EPROM, EEPROM, and flash memory devices, and the like. The device 500 may further include an input interface 508 and an output interface 510. The input interface 508 is configured to receive input signals and data. The output interface 510 is configured to send output signals and data.

[0066] The computer program may include instructions executable by a computer for causing the processor 504 of the device 500 to execute the method for longitudinal control of an autonomous vehicle of the present disclosure. The program may be recorded on any data storage medium including a memory. For example, the program may be implemented in digital electronic circuits, or in computer hardware, firmware, software, or in combinations thereof. The process / method steps described in the present disclosure may be executed by a programmable processor executing program instructions to perform the method, steps, operations by operating on input data and generating outputs.

[0067] Embodiments of the present disclosure may be implemented in a computer-readable medium. The computer-readable medium may store a computer program including instructions. In one example aspect, when the instructions are executed, they may cause at least one processor to: determine that there is at least one potential cut-in vehicle in front of the autonomous vehicle based on driving environment perception data; obtain a cut-in probability of each potential cut-in vehicle among the at least one potential cut-in vehicle; select at least one target vehicle based on the cut-in probability; and calculate an expected following distance for the at least one target vehicle.

[0068] Embodiments of the present disclosure may be implemented in a computer program product. The computer program product may include instructions. In one exemplary aspect, when the instructions are executed, they may cause a processor of a computing device to: determine that there is at least one potential cut-in vehicle in front of the autonomous vehicle based on driving environment perception data; obtain a cut-in probability of each potential cut-in vehicle among the at least one potential cut-in vehicle; select at least one target vehicle based on the cut-in probability; and calculate an expected following distance for the at least one target vehicle.

[0069] In addition to what has been described herein, various modifications may be made to the disclosed embodiments and implementations of the present invention without departing from the scope thereof. Accordingly, the description and examples herein should be construed as illustrative rather than limiting. The scope of the present invention should be measured only by reference to the claims.

Claims

1. A method for longitudinal control of an autonomous vehicle, comprising: Based on driving environment perception data, determining that there is at least one potential cut-in vehicle in front of the autonomous vehicle; Obtaining a cut-in probability for each of the at least one potential cut-in vehicle; Based on the cut-in probability, selecting at least one target vehicle; And Calculating an expected following distance for the at least one target vehicle.

2. The method according to claim 1, further comprising: The cut-in probability is generated by a probability prediction model based on the driving environment perception data.

3. The method according to claim 1, further comprising: Providing information related to the at least one selected target vehicle and the calculated expected following distance to a vehicle control unit of the autonomous vehicle for longitudinal control of the autonomous vehicle.

4. The method according to claim 1, further comprising: Determining a maximum probability threshold and a minimum probability threshold associated with the cut-in probability.

5. The method according to claim 4, further comprising: Based on the driving environment perception data, determining whether there is a current following vehicle in the current lane where the autonomous vehicle is located.

6. The method according to claim 5, wherein, In response to determining that there is no current following vehicle, the selecting of at least one target vehicle includes: Selecting the vehicle with the maximum cut-in probability among the at least one potential cut-in vehicle as the target vehicle, and wherein the expected following distance represents the expected distance between the autonomous vehicle and the target vehicle.

7. The method according to claim 6, wherein, The calculating of the expected following distance includes: Comparing the cut-in probability of the target vehicle with the maximum probability threshold and / or the minimum probability threshold; and Calculating the expected following distance based on the result of the comparison.

8. The method according to claim 7, wherein: If the cut-in probability of the target vehicle falls within a threshold interval defined by the minimum probability threshold and the maximum probability threshold, calculating the expected following distance based on the normalized cut-in probability of the target vehicle and a single-target following distance for the target vehicle; or If the cut-in probability of the target vehicle is greater than the maximum probability threshold, calculating the expected following distance based on the single-target following distance for the target vehicle.

9. The method according to claim 8, wherein, The normalized cut-in probability of the target vehicle is calculated by performing linear interpolation on the cut-in probability of the target vehicle within a threshold interval defined by the maximum probability threshold and the minimum probability threshold; and The single-target following distance for the target vehicle is calculated based on the time headway for the target vehicle and the current speed of the autonomous vehicle.

10. The method according to claim 5, wherein In response to determining that there is a current following vehicle, the selecting of at least one target vehicle includes: Selecting the current following vehicle, a first potential cut-in vehicle, and a second potential cut-in vehicle as target vehicles, and wherein the expected following distance represents the expected distance between the autonomous vehicle and the current following vehicle.

11. The method according to claim 10, wherein, The first potential cut-in vehicle is on the first adjacent lane of the current lane and is located between the autonomous vehicle and the current following vehicle; and The second potential cut-in vehicle is on the second adjacent lane of the current lane and is located between the autonomous vehicle and the first potential cut-in vehicle.

12. The method according to claim 11, wherein The first potential cut-in vehicle is the vehicle with the highest cut-in probability among multiple potential cut-in vehicles on the first adjacent lane; or The second potential cut-in vehicle is the vehicle with the highest cut-in probability among multiple potential cut-in vehicles on the second adjacent lane.

13. The method according to claim 12, wherein, The calculating of the expected following distance includes: Comparing the cut-in probabilities of the first potential cut-in vehicle and the second potential cut-in vehicle with the maximum probability threshold and / or the minimum probability threshold respectively; and Calculating the expected following distance based on the result of the comparison.

14. The method according to claim 12, wherein If the cut-in probabilities of the first potential cut-in vehicle and the second potential cut-in vehicle both fall within the threshold interval defined by the minimum probability threshold and the maximum probability threshold, then calculate the expected following distance based on the single-target following distance for the current following vehicle, the normalized cut-in probability of the first potential cut-in vehicle, the normalized cut-in probability of the second potential cut-in vehicle, the first incremental distance for the first potential cut-in vehicle, and the second incremental distance for the second potential cut-in vehicle.

15. The method according to claim 12, wherein If the cut-in probability of the first potential cut-in vehicle is greater than the maximum probability threshold, then calculate the expected following distance based on the single-target following distance for the first potential cut-in vehicle, the normalized cut-in probability of the second potential cut-in vehicle, and the second incremental distance for the second potential cut-in vehicle.

16. The method according to claim 12, wherein If the cut-in probability of the second potential cut-in vehicle is greater than the maximum probability threshold, then calculate the expected following distance based on the single-target following distance for the second potential cut-in vehicle.

17. The method according to claim 12, wherein If the cut-in probabilities of the first potential cut-in vehicle and the second potential cut-in vehicle are both less than the minimum probability threshold, then calculate the expected following distance based on the single-target following distance for the current following vehicle.

18. An apparatus for longitudinal control of an autonomous vehicle, comprising: A memory; A processor coupled to the memory, the processor being configured to execute the method according to any one of claims 1-17.

19. A computer-readable medium storing a computer program including instructions that, when executed by a processor, cause the processor to be configured to execute the method according to any one of claims 1-17.

20. A computer program product comprising instructions which, when executed by a processor of a computing device, cause the processor to be configured to perform the method according to any one of claims 1-17.

Images