Open Space Path Planning Using Inverse Reinforcement Learning

Abstract

In one embodiment, a method determines a route from a first location of an autonomous vehicle (ADV) in an open space to a second location, the first location being a current location of the ADV. The method determines an objective function based on the route, the objective function having a set of costs for maneuvering the ADV from the first location to the second location. The method determines environmental conditions of the open space and uses the environmental conditions to determine a set of weights, each weight to be applied to a corresponding cost of the objective function. The method optimizes the objective function based on one or more constraints so that an output of the objective function is minimized when the one or more constraints are satisfied, and generates a path trajectory using the optimized objective function to automatically control the ADV according to the path trajectory.

Description

Technical Field

[0001] Embodiments of the present disclosure generally relate to operating an autonomous vehicle. More specifically, embodiments of the present disclosure relate to open space path planning using inverse reinforcement learning. Background Art

[0002] A vehicle operating in an automated driving mode (e.g., without human control) can relieve occupants, particularly the driver, from some driving-related responsibilities. When operating in an automated driving mode, the vehicle can navigate to various locations using onboard sensors, allowing the vehicle to travel with minimal human interaction or in some cases without any passengers.

[0003] Motion planning and control are key operations in autonomous driving. In addition, trajectory planning is crucial for operating an autonomous vehicle (ADV). An open space planner can be utilized to optimize a trajectory based on various constraints. Based on the optimized trajectory, the ADV can generate commands to control the ADV along the path. However, in some cases, the trajectory generated by the planner often lacks sufficient smoothness required for autonomous driving. As a result, certain maneuvers performed by the ADV may be abrupt and less desirable to passengers. Summary of the invention

[0004] According to one aspect of the present disclosure, there is provided a computer-implemented method performed by an autonomous vehicle, the method comprising:

[0005] determining a route from a first location of the autonomous vehicle to a second location within an open space, the first location being a current location of the autonomous vehicle;

[0006] determining an objective function based on the route, the objective function having a set of costs for maneuvering the autonomous vehicle from the first location to the second location;

[0007] determining a set of weights based on one or more environmental conditions of the open space, each weight to be applied to a corresponding cost of the objective function;

[0008] Optimizing the objective function according to one or more constraints so that the output of the objective function is minimized when the one or more constraints are satisfied; and

[0009] A path trajectory is generated using the optimized objective function to automatically control the autonomous driving vehicle according to the path trajectory.

[0010] According to another aspect of the present disclosure, there is provided a non-transitory machine-readable medium having instructions stored therein, the instructions, when executed by a processor, causing the processor to perform operations, the operations comprising:

[0011] determining a route from a first location of the autonomous vehicle in an open space to a second location, the first location being a current location of the autonomous vehicle;

[0012] determining an objective function based on the route, the objective function having a set of costs for maneuvering the autonomous vehicle from the first location to the second location;

[0013] determining a set of weights based on one or more environmental conditions of the open space, each weight to be applied to a corresponding cost of the objective function;

[0014] Optimizing the objective function according to one or more constraints so that the output of the objective function is minimized when the one or more constraints are satisfied; and

[0015] A path trajectory is generated using the optimized objective function to automatically control the autonomous driving vehicle according to the path trajectory.

[0016] According to yet another aspect of the present disclosure, there is provided a data processing system, comprising:

[0017] Processor; and

[0018] a memory coupled to the processor to store instructions that, when executed by the processor, cause the processor to perform operations comprising:

[0019] determining a route from a first location of the autonomous vehicle in an open space to a second location, the first location being a current location of the autonomous vehicle;

[0020] determining an objective function based on the route, the objective function having a set of costs for maneuvering the autonomous vehicle from the first location to the second location;

[0021] determining a set of weights based on one or more environmental conditions of the open space, each weight to be applied to a corresponding cost of the objective function;

[0022] Optimizing the objective function according to one or more constraints so that the output of the objective function is minimized when the one or more constraints are satisfied; and

[0023] A path trajectory is generated using the optimized objective function to automatically control the autonomous driving vehicle according to the path trajectory. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Various aspects are shown by way of example and not limitation in the figures of the accompanying drawings, in which like reference numerals indicate similar elements. It should be noted that reference numerals for "an" or "one" aspect of the present disclosure are not necessarily the same aspect, and they are intended to refer to at least one. Moreover, for the sake of brevity and to reduce the total number of drawings, a given drawing may be used to illustrate features of more than one aspect, and not all elements in a drawing are required for a given aspect.

[0025] Figure 1 is a block diagram illustrating a networked system according to one embodiment.

[0026] Figure 2 is a block diagram illustrating an example of an autonomous driving vehicle according to one embodiment.

[0027] FIG. 3A to FIG. 3B is a block diagram illustrating an example of a perception and planning system for use with an autonomous vehicle, according to one embodiment.

[0028] Figure 4 A block diagram showing a system architecture for autonomous driving according to one embodiment.

[0029] Figure 5 is a flow chart of a process for dynamically determining one or more weights for open space path planning based on environmental conditions, according to one embodiment.

[0030] Fig. 6A and 6B is a stage illustrating the difference between static weights and weights dynamically determined based on environmental conditions according to another embodiment. DETAILED DESCRIPTION

[0031] Several embodiments of the present disclosure are now explained with reference to the accompanying drawings. Whenever the shapes, relative positions and other aspects of the components described in a given aspect are not clearly defined, the scope of the present disclosure herein is not limited to the parts shown, but is for illustrative purposes only. Moreover, although many details are set forth, it should be understood that some aspects can be implemented without these details. In other cases, well-known circuits, structures and technologies are not shown in detail to avoid confusing the understanding of this description. In addition, unless the meaning is obviously contrary, all ranges listed herein are deemed to include the endpoints of each range.

[0032] References in this specification to "one embodiment" or "an embodiment" mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the present disclosure. The appearances of the phrase "in one embodiment" in various places in this specification are not necessarily all referring to the same embodiment.

[0033] The present disclosure addresses the problem of improving the quality of path planning by adapting a path planning system based on environmental conditions to control an ADV to drive as if it were manually controlled by a human operator (e.g., when in the driver's seat). As described herein, a path planning system can create an objective function with one or more costs (or cost functions) for automatically controlling an ADV along a path or route. Weights can be associated and applied to each cost (e.g., multiplication) to modify the impact of each cost on the overall computational cost. In conventional planning systems, these weights are static (e.g., do not change predetermined values).

[0034] However, the solution proposed in the current disclosure provides an open space path planning method that uses a reward function to adjust (or modify) at least some weights based on driving conditions. Specifically, the reward function is created by observing the behavior of (professional) drivers during different driving goals according to an inverse reinforcement learning (IRL) framework. When creating the objective function, the planning system determines environmental conditions (e.g., images captured by a camera of the ADV) and applies the conditions to the reward function to generate weights (or different weight values). Therefore, the generated weights (when applied to the objective function) reflect the "human-like" control of the ADV, thereby providing a more pleasant driving experience for passengers.

[0035] According to some embodiments, a computer-implemented method performed by an ADV is disclosed. The method includes determining a route from a first position of the ADV in an open space to a second position, the first position being the current position of the ADV. The method determines an objective function based on the route, the objective function having a cost set for maneuvering the ADV from the first position to the second position. The method determines environmental conditions (or parameters) of the open (or driving) space and uses these parameters to determine a set of weights, each weight being applied to a corresponding cost of the objective function. The method optimizes the objective function based on one or more constraints so that the output of the objective function is minimized while satisfying the one or more constraints. The method generates a path trajectory using the optimized objective function to automatically control the ADV according to the path trajectory.

[0036] In one embodiment, using the environmental conditions to determine the set of weights includes performing a table lookup in a table associating environmental conditions with weights, the predefined table being an inverse reinforcement learning (IRL) model of weights representing the behavior of an expert driver with respect to the environmental conditions.

[0037] In another embodiment, determining the environmental condition of the open space includes analyzing images captured by one or more cameras of the ADV to identify the environmental condition. In one embodiment, the environmental condition includes the position of one or more objects captured within the image relative to the current position of the ADV. In some embodiments, the environmental condition includes road features of the open space captured within the image.

[0038] In one embodiment, the method further includes: controlling the ADV to maneuver from the first position to the second position along the generated path trajectory. In some embodiments, the route is a first route, the objective function is a first objective function, the open space is a first open space, the weight set is a first weight set, and the cost set is a first cost set. The method further includes: determining a second route from a second position of the ADV within a second open space to a third position, the second position being the current position of the ADV; determining a second objective function based on the second route, the second objective function including a second cost set for maneuvering from the second position to the third position; determining environmental conditions of the second open space; and using the environmental conditions of the second open space to determine a second weight set, each weight to be applied to a corresponding cost of the second objective function. In some embodiments, at least one weight in the second weight set is different from the corresponding weight in the first weight set.

[0039] In another embodiment of the present disclosure, a non-transitory machine-readable medium and a data processing system perform at least some of the processes described herein.

[0040] Figure 1 is a block diagram showing a network configuration of an autonomous driving vehicle according to one embodiment of the present disclosure. Figure 1 , the network configuration 100 includes an autonomous driving vehicle (ADV) 101 that can be communicatively connected to one or more servers 103 to 104 via a network 102. Although one autonomous driving vehicle is shown, multiple autonomous driving vehicles can be connected to each other and / or to servers 103 to 104 via the network 102. The network 102 can be any type of network, for example, a wired or wireless local area network (LAN), a wide area network (WAN) such as the Internet, a cellular network, a satellite network, or a combination thereof. Servers 103 to 104 can be any type of server or server cluster, such as a web or cloud server, an application server, a back-end server, or a combination thereof. Servers 103 to 104 can be data analysis servers, content servers, traffic information servers, maps and points of interest (MPOI) servers, or location servers, etc.

[0041] An autonomous vehicle refers to a vehicle that is configurable to be in an autonomous driving mode in which the vehicle navigates through an environment with little or no input from a driver. Such an autonomous vehicle may include a sensor system having one or more sensors configured to detect information related to the environment in which the vehicle operates. The vehicle and its associated controller use the detected information to navigate through the environment. The autonomous vehicle 101 may operate in a manual mode, in a fully autonomous driving mode, or in a partially autonomous driving mode.

[0042] In one embodiment, the autonomous vehicle 101 includes, but is not limited to, a perception and planning system 110, a vehicle control system 111, a wireless communication system 112, a user interface system 113, and a sensor system 115. The autonomous vehicle 101 may also include certain common components included in an ordinary vehicle, such as: an engine, wheels, a steering wheel, a transmission, etc., and the components may be controlled by the vehicle control system 111 and / or the perception and planning system 110 using a variety of communication signals and / or commands, such as acceleration signals or commands, deceleration signals or commands, steering signals or commands, braking signals or commands, etc.

[0043] Components 110 to 115 may be communicatively coupled to each other via an interconnect, a bus, a network, or a combination thereof. For example, components 110 to 115 may be communicatively coupled to each other via a controller area network (CAN) bus. The CAN bus is a vehicle bus standard designed to allow microcontrollers and devices to communicate with each other in applications without a host. It is a message-based protocol originally designed for multiplexed electrical wiring within an automobile, but is also used in many other environments.

[0044] Reference now Figure 2 In one embodiment, the sensor system 115 includes, but is not limited to, one or more cameras 211, a global positioning system (GPS) unit 212, an inertial measurement unit (IMU) 213, a radar unit 214, and a light detection and ranging (LIDAR) unit 215. The GPS system 212 may include a transceiver that is operable to provide information about the location of the autonomous vehicle. The IMU unit 213 may sense the position and orientation changes of the autonomous vehicle based on inertial acceleration. The radar unit 214 may represent a system that uses radio signals to sense objects within the local environment of the autonomous vehicle. In some embodiments, in addition to sensing objects, the radar unit 214 may additionally sense the speed and / or heading of the object. The LIDAR unit 215 may use lasers to sense objects in the environment of the autonomous vehicle. The LIDAR unit 215 may also include, among other system components, one or more laser sources, a laser scanner, and one or more detectors. The camera 211 may include one or more devices for capturing images of the environment surrounding the autonomous vehicle. The camera 211 may be a still camera and / or a video camera. The camera may be mechanically movable, for example, by mounting the camera on a rotating and / or tilting platform.

[0045] The sensor system 115 may also include other sensors, such as: a sonar sensor, an infrared sensor, a steering sensor, a throttle sensor, a brake sensor, and an audio sensor (e.g., a microphone). The audio sensor may be configured to collect sound from the environment around the autonomous vehicle. The steering sensor may be configured to sense the steering angle of the steering wheel, the wheels of the vehicle, or a combination thereof. The throttle sensor and the brake sensor sense the throttle position and the brake position of the vehicle, respectively. In some cases, the throttle sensor and the brake sensor may be integrated into an integrated throttle / brake sensor.

[0046] In one embodiment, the vehicle control system 111 includes, but is not limited to, a steering unit 201, a throttle unit 202 (also referred to as an acceleration unit), and a brake unit 203. The steering unit 201 is used to adjust the direction or forward direction of the vehicle. The throttle unit 202 is used to control the speed of the motor or engine, which in turn controls the speed and acceleration of the vehicle. The brake unit 203 decelerates the vehicle by providing friction to decelerate the wheels or tires of the vehicle. It should be noted that, as Figure 2 The components shown may be implemented in hardware, software or a combination thereof.

[0047] Return to reference Figure 1 , the wireless communication system 112 allows communication between the autonomous driving vehicle 101 and external systems such as devices, sensors, other vehicles, etc. For example, the wireless communication system 112 can communicate directly with one or more devices wirelessly, or communicate wirelessly via a communication network, such as communicating with servers 103 to 104 through the network 102. The wireless communication system 112 can use any cellular communication network or wireless local area network (WLAN), for example, using WiFi, to communicate with another component or system. The wireless communication system 112 can communicate directly with a device (e.g., a passenger's mobile device, a display device, a speaker in the vehicle 101) using, for example, an infrared link, Bluetooth, etc. The user interface system 113 can be part of a peripheral device implemented in the vehicle 101, including, for example, keywords, a touch screen display device, a microphone and a speaker, etc.

[0048] Some or all of the functions of the autonomous vehicle 101 may be controlled or managed by the perception and planning system 110, particularly when operating in an autonomous mode. The perception and planning system 110 includes the necessary hardware (e.g., processor, memory, storage) and software (e.g., operating system, planning and routing programs) to receive information from the sensor system 115, the control system 111, the wireless communication system 112, and / or the user interface system 113, process the received information, plan a route or path from a starting point to a destination point, and then drive the vehicle 101 based on the planning and control information. Alternatively, the perception and planning system 110 may be integrated with the vehicle control system 111.

[0049] For example, a user as a passenger may specify a starting location and a destination for a trip, for example, via a user interface. The perception and planning system 110 obtains trip-related data. For example, the perception and planning system 110 may obtain location and route information from an MPOI server, which may be part of the servers 103 to 104. The location server provides location services, and the MPOI server provides map services and POIs for certain locations. Alternatively, such location and MPOI information may be locally cached in a permanent storage device of the perception and planning system 110.

[0050] As the autonomous vehicle 101 moves along the route, the perception and planning system 110 may also obtain real-time traffic information from a traffic information system or server (TIS). It should be noted that the servers 103 to 104 may be operated by a third-party entity. Alternatively, the functionality of the servers 103 to 104 may be integrated with the perception and planning system 110. Based on the real-time traffic information, MPOI information, and location information, as well as real-time local environmental data (e.g., obstacles, objects, nearby vehicles) detected or sensed by the sensor system 115, the perception and planning system 110 may plan an optimal route and drive the vehicle 101 according to the planned route, for example, via the control system 111, to safely and efficiently reach a designated destination.

[0051] The server 103 may be a data analysis system, thereby performing data analysis services for various clients. In one embodiment, the data analysis system 103 includes a data collector 121 and a machine learning engine 122. The data collector 121 collects driving statistics 123 from various vehicles (autonomous driving vehicles or conventional vehicles driven by human drivers). The driving statistics 123 include information indicating the driving instructions issued (e.g., throttle, brake, steering instructions) and the responses of the vehicle (e.g., speed, acceleration, deceleration, direction) captured by the vehicle's sensors at different points in time. The driving statistics 123 may also include information describing the driving environment at different points in time, such as routes (including starting and destination locations), MPOI, road conditions, weather conditions, etc.

[0052] Based on the driving statistics 123, the machine learning engine 122 generates or trains a set of rules, algorithms, and / or predictive models 124 for various purposes. In one embodiment, the model 124 may include an inverse reinforcement learning (IRL) model of the present disclosure. Specifically, the IRL model may be a reward function derived (or learned) using an IRL algorithm, wherein the model indicates a goal or value of an agent (or professional driver) based on observed behavior. In particular, (e.g., in a laboratory setting) the IRL algorithm may obtain inputs such as a driver's path trajectory (or action) and input states (e.g., environmental conditions, such as traffic conditions), and the algorithm may output (or learn) one or more weights representing the driver's observed behavior as an IRL model. In one embodiment, the model may be a data structure containing a lookup table that associates the weights output by the IRL algorithm with one or more environmental conditions. More information about IRL models will be described herein.

[0053] In one embodiment, the algorithms and / or models (e.g., IRL models) 124 can then be uploaded to the ADV for use in real time during autonomous driving (e.g., by data stored in the ADV). Figure 3A In another embodiment, the model may be uploaded periodically (eg, once a day) so that the model is updated periodically as needed.

[0054] Figure 3A and Figure 3B is a block diagram illustrating an example of a perception and planning system for use with an autonomous vehicle according to one embodiment. System 300 may be implemented as Figure 1 A portion of an autonomous vehicle 101 includes, but is not limited to, a perception and planning system 110, a control system 111, and a sensor system 115. FIG. 3A to FIG. 3B The perception and planning system 110 includes but is not limited to a positioning module 301 , a perception module 302 , a prediction module 303 , a decision module 304 , a planning module 305 including an open space planning module 308 , a control module 306 and a route scheduling module 307 .

[0055] Some or all of the modules 301 to 308 may be implemented in software, hardware, or a combination thereof. For example, these modules may be installed in permanent storage 352, loaded into memory 351, and executed by one or more processors (not shown). It should be noted that some or all of these modules may be communicatively coupled to Figure 2 Some or all modules of the vehicle control system 111 of the embodiment of the present invention may be integrated with them. Some of the modules 301 to 308 may be integrated together into an integrated module.

[0056] The positioning module 301 determines the current location of the autonomous vehicle 300 (e.g., using the GPS unit 212) and manages any data related to the user's itinerary or route. The positioning module 301 (also referred to as the map and route module) manages any data related to the user's itinerary or route. The user can log in, for example, via a user interface and specify the starting location and destination of the itinerary. The positioning module 301 communicates with other components of the autonomous vehicle 300, such as the map and route information 311, to obtain itinerary related data. For example, the positioning module 301 can obtain location and route information from a location server and a map and POI (MPOI) server. The location server provides location services, and the MPOI server provides map services and POIs at certain locations, which can be cached as part of the map and route information 311. When the autonomous vehicle 300 moves along the route, the positioning module 301 can also obtain real-time traffic information from a traffic information system or server.

[0057] Based on the sensor data provided by the sensor system 115 and the positioning information obtained by the positioning module 301, the perception module 302 determines the perception of the surrounding environment. The perception information may represent what a normal driver would perceive around the vehicle being driven by the driver as environmental condition data 313. The environmental condition data 313 may include, for example, lane configuration in the form of objects, traffic light signals, the relative position of another vehicle, pedestrians, buildings, crosswalks or other traffic-related signs (e.g., stop signs, yield signs), etc. The lane configuration includes information describing one or more lanes, such as, for example, the shape of the lane (e.g., straight or curved), the width of the lane, the number of lanes in the road, one-way or two-way lanes, merging or splitting lanes, exit lanes, etc.

[0058] The perception module 302 may include a computer vision system or the functionality of a computer vision system to process and analyze images captured by one or more cameras to identify objects and / or features in the autonomous vehicle environment as environmental condition data 313. Objects may include traffic signals, road boundaries, other vehicles, pedestrians, and / or obstacles, etc. In one embodiment, the data 313 may indicate spatial characteristics of the object. For example, the data may indicate the position (relative to the current position of the ADV 101) of one or more objects (e.g., other vehicles) within the captured image. In another embodiment, the data 313 may indicate characteristics of the object, such as (estimates of) the speed and acceleration of the identified vehicle. As another example, the feature may be a road feature (e.g., lane, curvature of the road, etc.) within the open driving space where the ADV is located. The computer vision system may use object recognition algorithms, video tracking, and other computer vision techniques. In some embodiments, the computer vision system may map the environment, track objects, and estimate the speed of objects, etc. The perception module 302 may also detect objects based on other sensor data provided by other sensors such as radar and / or LIDAR.

[0059] In one embodiment, driving / traffic rules 312 may include traffic information, such as construction traffic signs near the ADV, so that the ADV may avoid lanes with such construction signs. In one embodiment, rules 312 may be part of environmental condition data 313.

[0060] For each object, the prediction module 303 predicts how the object will behave in this situation. The prediction is performed based on the perception data, which perceives the driving environment at a point in time considering a set of map / route information 311 and traffic rules 312. For example, if the object is a vehicle in the opposite direction and the current driving environment includes an intersection, the prediction module 303 will predict whether the vehicle is likely to move straight ahead or turn. If the perception data indicates that there is no traffic light at the intersection, the prediction module 303 may predict that the vehicle may need to come to a complete stop before entering the intersection. If the perception data indicates that the vehicle is currently in a left-turn only lane or a right-turn only lane, the prediction module 303 may predict that the vehicle will be more likely to turn left or right, respectively.

[0061] For each object, the decision module 304 makes a decision about how to handle the object. For example, for a particular object (e.g., another vehicle in a crossing path) and metadata describing the object (e.g., speed, direction, turn angle), the decision module 304 decides how to encounter the object (e.g., pass, yield, stop, overtake). The decision module 304 may make such decisions based on a rule set such as traffic rules or driving rules 312, which may be stored in the persistent storage device 352.

[0062] The route arrangement module 307 is configured to provide one or more routes or paths from the starting point to the destination point. For a given trip from the starting location to the destination location, such as a given trip received from a user, the route arrangement module 307 obtains the route and map information 311 and determines all possible routes or paths from the starting location to the destination location. The route arrangement module 307 can generate a reference line in the form of a topographic map, which determines each route from the starting location to the destination location. The reference line refers to an ideal route or path that is not interfered by any other factors such as other vehicles, obstacles or traffic conditions. That is, if there are no other vehicles, pedestrians or obstacles on the road, the ADV should follow the reference line accurately or closely. Then, the topographic map is provided to the decision module 304 and / or the planning module 305. The decision module 304 and / or the planning module 305 examine all possible routes to select and change one of the best routes based on other data provided by other modules, wherein the other data is such as traffic conditions from the positioning module 301, the driving environment perceived by the perception module 302, and the traffic conditions predicted by the prediction module 303. Depending on the particular driving circumstances at a point in time, the actual path or route used to control the ADV may be close to or different from the reference line provided by the routing module 307 .

[0063] Based on the decision for each of the perceived objects, the planning module 305 plans a path or route and driving parameters (e.g., distance, speed, and / or turning angle) for the autonomous vehicle using the reference line provided by the routing module 307 as a basis. In other words, for a given object, the decision module 304 decides what to do with the object, and the planning module 305 determines how to do it. For example, for a given object, the decision module 304 may decide to overtake the object, and the planning module 305 may determine whether to overtake on the left or right side of the object. More information about the planning module 305 will be described herein. In one embodiment, planning and control data is generated by the planning module 305, including information describing how the vehicle 300 will move in the next movement cycle (e.g., the next route / path segment). For example, the planning and control data may indicate that the vehicle 300 moves at a speed of 30 miles per hour (mph) for 10 meters and then changes to the right lane at a speed of 25 mph.

[0064] Based on the planning and control data, the control module 306 controls and drives the autonomous vehicle according to the route or path defined by the planning and control data by sending appropriate commands or signals to the vehicle control system 111. The planning and control data includes sufficient information to drive the vehicle from a first point on the route or path to a second point using appropriate vehicle settings or driving parameters (e.g., throttle, brake, steering commands) at different points in time along the route or path.

[0065] In one embodiment, the planning phase is performed in multiple planning cycles (also referred to as driving cycles), for example, in each cycle with a time interval of 100 milliseconds (ms). For each of the planning cycles or driving cycles, one or more control commands will be issued based on the planning and control data. That is, for every 100ms, the planning module 305 plans the next route segment or path segment, for example, including the target position and the time required for the ADV to reach the target position. Alternatively, the planning module 305 can also specify a specific speed, direction and / or steering angle, etc. In one embodiment, the planning module 305 plans the route segment or path segment for the next predetermined period of time (such as 5 seconds). For each planning cycle, the planning module 305 plans the target position for the current cycle (e.g., the next 5 seconds) based on the target position planned in the previous cycle. The control module 306 then generates one or more control commands (e.g., throttle, brake, steering control commands) based on the planning and control data of the current cycle.

[0066] It should be noted that the decision module 304 and the planning module 305 can be integrated into an integrated module. The decision module 304 / planning module 305 may include a navigation system or the functionality of a navigation system to determine a driving path for an autonomous vehicle. For example, the navigation system may determine a series of speeds and directions of travel for influencing the autonomous vehicle to move along a path that substantially avoids perceived obstacles while allowing the autonomous vehicle to move along a roadway-based path to a final destination. The destination may be set based on user input via the user interface system 113. The navigation system may dynamically update the driving path while the autonomous vehicle is running. The navigation system may combine data from a GPS system and one or more maps to determine a driving path for the autonomous vehicle.

[0067] In one embodiment, the open space planning module 308 may include an open space planner that may be configured to generate a path trajectory for the ADV 101 in an open space without reference lines or lanes to follow. Examples of open spaces include parking lots or roads where vehicles perform parallel parking, U-turns, or three-point turns. In one embodiment, the open space planner may also generate a path trajectory with one or more reference lines.

[0068] In one embodiment, the planner may determine an objective function including one or more costs (or cost functions), such as included in cost function 315, associated with navigating the ADV along the route (as determined by routing module 307). In one example, the objective function may be

[0069] J=w1*Cost1+w2*Cost2+w3*Cost3+…+w n *Costn

[0070] Where J is the calculated total cost, w n Corresponds to Cost n The weight of (e.g., can be a value, such as a value from 0 to 1). In one embodiment, the objective function may include one or more costs. The cost may be optimized by minimizing the objective function J. The cost may include at least one of the following: cross-track error (compensation for the distance between the ADV and the target track), forward direction error (compensation for the error between the ADV forward direction and the target track direction at a point), speed cost (compensation for speed change), steering cost (compensation for steering change), acceleration cost (compensation for acceleration change), steering rate of change (compensation for how fast the steering changes), braking cost (compensation for braking), acceleration rate of change (compensation for how fast the acceleration changes), distance to the target location (e.g., the distance between the ADV and an identified obstacle (such as another vehicle), or the distance between the ADV and the destination), curvature of the route (or reference line), time efficiency along the route (e.g., how long it takes to reach the destination location from a location), passenger comfort (e.g., based on sudden movements caused by acceleration and deceleration or braking), minimum curvature (e.g., the distance that the ADV may deviate from the route or reference line), and minimum curvature rate of change. In one embodiment, control commands such as throttle, steering, and braking (eg, commands provided by control module 306 ) can be generated, as described herein, to optimally track a path trajectory while taking into account the cost of optimizing an objective function.

[0071] In one embodiment, the open space planning module 308 is configured to determine at least some weights w that will be applied to the corresponding costs of the objective function. Specifically, the module can obtain environmental condition data 313 and apply the data to the IRL model to generate weights. For example, the module can use the data 313 to perform a table lookup in the IRL model to identify and select at least some of the weights associated with the data. More information about determining weights is described herein.

[0072] In one embodiment, the open space planning module 308 uses an optimization model (e.g., a quadratic programming (QP) optimizer) to optimize an objective function (in which one or more determined weights have been applied to one or more corresponding costs) to generate an optimal path trajectory based on one or more constraints. Constraints may include inequality, equality, and boundary constraints. In one embodiment, constraints may include features associated with the ADV, such as a velocity (e.g., speed) range of the ADV (e.g., how fast the ADV can travel), an acceleration range of the ADV (e.g., how fast the ADV can accelerate and / or decelerate), a distance in front of the ADV (e.g., a distance between the ADV and an obstacle (e.g., another vehicle located in front of the ADV)). In another embodiment, the constraints may be related to the open space, such as a speed limit within the open space, a distance between the ADV and a road feature (e.g., such as a curb), a threshold curvature of a road traversed along the route, etc.

[0073] As described herein, the control module 306 uses the optimal path trajectory to control (or steer) the ADV (e.g., issue control commands) along the path trajectory. In one embodiment, the optimal path trajectory may be the same as the reference line or route determined by the routing module 307. In another embodiment, the trajectory may be (slightly) different from the determined route. Specifically, the trajectory may follow the same "regular" path as the route, meaning that the trajectory may make at least some of the same maneuvers (e.g., turn left at 0.3 miles).

[0074] Figure 4 400 is a block diagram showing a system architecture for autonomous driving according to one embodiment. Figure 3A and 3B The system architecture of the autonomous driving system is shown in Figure 2. Figure 4, the system architecture 400 includes, but is not limited to, an application layer 401, a planning and control (PNC) layer 402, a perception layer 403, a driver layer 404, a firmware layer 405, and a hardware layer 406. The application layer 401 may include a user interface or configuration application that interacts with a user or passenger of the autonomous vehicle, such as functionality associated with the user interface system 113. The PNC layer 402 may include at least the functionality of the planning module 305 and the control module 306. The perception layer 403 may include at least the functionality of the perception module 302. In one embodiment, there is an additional layer that includes the functionality of the prediction module 303 and / or the determination module 304. Alternatively, such functionality may be included in the PNC layer 402 and / or the perception layer 403. The system architecture 400 also includes a driver layer 404, a firmware layer 405, and a hardware layer 406. The firmware layer 405 may represent at least the functionality of the sensor system 115, which may be implemented in the form of a field programmable gate array (FPGA). The hardware layer 406 may represent hardware of the autonomous vehicle, such as the control system 111. The layers 401 to 403 may communicate with the firmware layer 405 and the hardware layer 406 via the device driver layer 404 .

[0075] Figure 5 4 is a flowchart of a process for dynamically determining one or more weights for open space path planning based on environmental conditions according to one embodiment. Specifically, the process determines (or creates) one or more weights (values) to apply to an objective function so that the ADV provides a "human-like" driving experience. The process can be performed by one or more processors (e.g., in hardware layer 406) of an ADV (e.g., ADV 101). In one embodiment, at least some of the operations described herein may be performed by one or more of the modules 301-308 of the perception and planning system 110, such as the planning module 305 (open space planning module 308).

[0076] Process 500 determines a route from a first location (e.g., current location) of an ADV to a second location within an open (driving) space (at block 501). As described herein, routing module 307 may determine (provide) a route from a starting location to a destination. For example, routing module 307 may receive route and map data based on a destination received from a user. In one embodiment, the determined route may be one of several route segments that integrate the entire route. In one embodiment, routing module 307 may generate a reference line in the form of a topographic map including the route, the starting location, and the destination.

[0077] The process 500 determines an objective function including one or more costs for maneuvering the ADV along the route from a first location to a second location based on the determined route (at block 502). For example, the open space planning module 308 may determine the one or more costs from the cost function 315 associated with the route. As described herein, the costs may include a distance to the target location, a minimum curvature, and a minimum rate of change of curvature.

[0078] Process 500 determines the environmental conditions of the open space (at box 503). As described herein, one or more cameras 211 can capture images containing objects within the camera's field of view. The computer vision system (of the perception module 302) can analyze the images to identify the environmental conditions contained therein. For example, the computer vision system can identify objects (e.g., other vehicles) within the open space. Specifically, the conditions can include position data (position) of the objects within the image relative to (the current position of) the ADV. For example, the position data can be a (e.g., two-dimensional) map of the open space around the ADV (the ADV is in the middle of the map). The map can include one or more areas, each area specifies a numerical value, where one value (e.g., "1") indicates that there is an object within the area, and another value (e.g., "0") indicates that the space is unobstructed or does not have any objects. In another embodiment, the conditions can include characteristics of the object (e.g., the speed at which the object moves relative to the ADV). In some embodiments, the conditions can include road features contained in the captured image (e.g., the number of lanes, road curvature, traffic signals, etc.). In one embodiment, the environmental conditions can be determined in real time (e.g., when the ADV is navigating in the open space).

[0079] The process 500 uses the environmental conditions to determine a set of weights, each weight to be applied to a corresponding cost of the objective function (at box 504). Specifically, the open space planning module 308 can apply the environmental conditions to one or more IRL models 314 to determine the set of weights. For example, as described herein, the model can be a (predefined) lookup table of reward functions associated with weights, and the weights represent the behavior of professional drivers with respect to environmental conditions. In this example, as described herein, the module 308 can use the determined environmental conditions to identify and select one or more weights, and for each cost of the objective function, perform a table lookup in a lookup table that associates the environmental condition with a weight for the one or more costs.

[0080] In one embodiment, each determined weight may be applied to (e.g., multiplied by) a corresponding cost. For example, the determined objective function may only include costs (e.g., without applying weights to them). The planning module 308 may multiply each cost by a corresponding weight (or weight value). In another embodiment, the module may apply weights by updating existing weights applied to costs. For example, the determined objective function may include applied weights. In one embodiment, the weights may be predefined weights. When applying the determined weights, the module may multiply the existing weights by the corresponding determined weights.

[0081] The process 500 optimizes (e.g., using a QP optimizer) an objective function (at box 505) based on one or more constraints so that the output of the objective function is minimized while satisfying the one or more constraints. For example, the optimized objective function may be a feasible solution such that no constraints are violated. In one embodiment, one or more constraints may be "soft" constraints that are allowed to be violated during the optimization process. The process 500 generates a path trajectory using the optimized objective function to automatically control the ADV according to the path trajectory (at box 506). In one embodiment, the path trajectory may be output (e.g., to the control module 306) to control (or manipulate) the ADV along the generated path trajectory from a current position to a desired position. As described herein, the path trajectory may be the same as or different from the determined route. For example, the path trajectory may be "smoother" than the route. For example, the trajectory may have fewer sudden movements than the determined route.

[0082] The process 500 determines whether the ADV is to travel to a new location (at decision block 507). As described herein, a route may be a route segment. Thus, the process may determine that the ADV is to navigate along another segment of the primary route. In another embodiment, the routing module 307 may receive data (e.g., map data) that the user wishes to travel from the current location to the new location. If the ADV is to travel to the new location, the process 500 returns to block 501 to repeat at least some of the operations of the process 500. As described herein, each time the process is repeated, one or more weights may be different (e.g., from previously determined weights).

[0083] Some embodiments perform variations of process 500. For example, specific operations of the process may not be performed in the exact order shown and described. Specific operations may not be performed in a continuous series of operations, some operations may be omitted, and different specific operations may be performed in different embodiments.

[0084] Fig. 6A and 6Bis a stage showing the difference between static weights and weights dynamically determined based on environmental conditions according to another embodiment. Each figure shows that ADV 601 dynamically determines the weights to be applied to the objective function based on changes in environmental conditions.

[0085] Fig. 6A An open (driving) space 600 is shown including an ADV 601 and three obstacles 602-604. In one embodiment, the ADV 601 may be Figure 1 ADV 601 may be similar to (or identical to) ADV 101. For example, ADV 601 may include at least some systems, such as perception and planning system 110. In some embodiments, open space 600 may be any space in which ADV 601 may travel, such as a parking lot. As another example, the open space may be a road (e.g., having one or more lanes). In one embodiment, an obstacle may be any object that the ADV avoids contacting (e.g., maintaining a distance greater than a threshold distance, such as a foot away) while navigating through the space. For example, obstacles may be other ADVs or vehicles. In another embodiment, at least some obstacles may be stationary obstacles (e.g., light poles). In some embodiments, at least some obstacles may be associated with driving / traffic rules (contained in rules 312), such as painted lines on a road.

[0086] As shown, ADV 601 is in the starting (first) position 705 within the open space 600, and the space also includes a destination position 606. Two path trajectories 607 and 608 are also shown. In one embodiment, the two trajectories may have been generated using the same objective function (e.g., with the same cost), but with different weights. Specifically, the objective function may include a turning cost, as described herein. The first path trajectory 607 is generated using an optimized objective function with static or predefined weights. For example, the weight applied to the turning cost may be low (e.g., 0.2), thereby allowing the ADV to make rapid turning changes. As a result, in order to avoid obstacle 604 and reach destination 606, the path trajectory 607 along which the ADV will move causes the ADV to turn sharply toward obstacle 602 and then go straight to the destination. However, due to the sharp turning motion, this trajectory may be undesirable for passengers.

[0087] On the other hand, second path trajectory 608 is generated using an optimized objective function with weights determined using an IRL model based on environmental conditions (e.g., the location of obstacles). In this case, the weights from the IRL model provide a more "human-like" response by increasing the weight for the steering cost (e.g., to 0.8) to more smoothly steer around obstacle 604. In one embodiment, the weight can be based on the location of at least one obstacle within open space 600, such as obstacle 604.

[0088] Figure 6B It is shown that the weights can be changed dynamically based on changes in environmental conditions. Specifically, the figure shows the next (second) route (or route segment) that ADV601 will travel. For example, the figure shows a (second) open space 700, including obstacles 701 and 702 that are not in the open space 600, and also including obstacles 603 and 604 in the open space 600. In one embodiment, since the obstacles have no effect on the generation of the path trajectory described herein, the obstacle 602 is not included in the open space 700 (shown as a dotted line).

[0089] In this figure, ADV 601 is in a starting position 606 ( Fig. 6A ), and space 700 has a (new) destination position 703. It is also shown that the same objective function (possibly with Fig. 6A 704 and 705. In particular, path trajectory 704 may have been generated using static weights, while trajectory 705 may have been generated using dynamically determined weights, as described herein. Again, trajectory 704 generated with static weights makes sharp motions (e.g., having a weight of 0.2 applied to steering costs). However, trajectory 705 has a smoother turn. The trajectory is generated using an optimized (second) objective function with steering weights determined from an IRL model based on environmental conditions as weights. In this case, the weights may be less than the previously determined weights (e.g., 0.6), thereby compensating less for steering costs than with higher weights in order to allow the ADV to make slightly steeper turns.

[0090] It should be noted that some or all of the components shown and described above may be implemented in software, hardware, or a combination thereof. For example, such components may be implemented as software installed and stored in a permanent storage device, which may be loaded into a memory by a processor (not shown) and executed in the memory to implement the processes or operations throughout the present application. Alternatively, such components may be implemented as executable code programmed or embedded into dedicated hardware (such as an integrated circuit (e.g., an application-specific integrated circuit or ASIC), a digital signal processor (DSP), or a field programmable gate array (FPGA)), which may be accessed via a corresponding driver and / or operating system from an application. In addition, such components may be implemented as specific hardware logic in a processor or processor core as part of an instruction set that may be accessed by a software component through one or more specific instructions.

[0091] Some portions of the foregoing detailed description have been presented in terms of algorithms and symbolic representations of operations on data bits within a computer memory. These algorithmic descriptions and representations are the means used by those skilled in the data processing arts to most effectively convey the substance of their work to others skilled in the art. An algorithm is generally considered herein to be a self-consistent sequence of operations leading to a desired result. These operations are those requiring physical manipulation of physical quantities.

[0092] It should be borne in mind, however, that all of these and similar terms are intended to be associated with the appropriate physical quantities and are merely convenient labels applied to these quantities. Unless otherwise expressly indicated in the above discussion, it should be understood that throughout the specification, discussions utilizing terms such as those set forth in the appended claims refer to the actions and processes of a computer system or similar electronic computing device that manipulates data represented as physical (electronic) quantities within the computer system's registers and memories and transforms the data into other data similarly represented as physical quantities within the computer system's memories or registers or other such information storage, transmission, or display devices.

[0093] Embodiments of the present disclosure also relate to devices for performing the operations herein. Such a computer program is stored in a non-transitory computer-readable medium. A machine-readable medium includes any mechanism for storing information in a form readable by a machine (e.g., a computer). For example, a machine-readable (e.g., computer-readable) medium includes a machine (e.g., computer) readable storage medium (e.g., a read-only memory ("ROM"), a random access memory ("RAM"), a magnetic disk storage medium, an optical storage medium, a flash memory device).

[0094] The processes or methods depicted in the foregoing figures may be performed by processing logic including hardware (e.g., circuitry, dedicated logic, etc.), software (e.g., embodied on a non-transitory computer-readable medium), or a combination of both. Although the processes or methods are described above in terms of some sequential operations, it should be understood that some of the operations may be performed in a different order. Furthermore, some operations may be performed in parallel rather than sequentially.

[0095] The embodiments of the present disclosure are not described with reference to any particular programming language. It will be appreciated that a variety of programming languages ​​may be used to implement the teachings of the embodiments of the present disclosure as described herein.

[0096] In the above description, embodiments of the present disclosure have been described with reference to specific exemplary embodiments of the present disclosure. It will be apparent that various modifications may be made to the present invention without departing from the broader spirit and scope of the present disclosure as set forth in the appended claims. Therefore, the present description and the accompanying drawings should be understood in an illustrative sense rather than a restrictive sense.

[0097] In some embodiments, the present disclosure may include language such as “at least one of [element A] and [element B].” The language may refer to one or more elements. For example, “at least one of A and B” may refer to “A,” “B,” or “A and B.” Specifically, “at least one of A and B” may refer to “at least one of A and at least one of B” or “at least one of A or B.” In some embodiments, the present disclosure may include language such as “[element A], [element B], and / or [element C].” The language may refer to any one of the elements or any combination thereof. For example, “A, B, and / or C” may refer to “A,” “B,” “C,” “A and B,” “A and C,” “B and C,” or “A, B, and C.”

Claims

1. A computer-implemented method performed by an autonomous vehicle, the method comprising: determining a first route from a first location of the autonomous vehicle to a second location within a first open space, the first location being a current location of the autonomous vehicle; determining a first objective function based on the first route, the first objective function having a first set of costs for maneuvering the autonomous vehicle from the first location to the second location; determining a first set of weights based on one or more environmental conditions of the first open space, each weight in the first set of weights to be applied to a corresponding cost of the first objective function; Optimizing the first objective function according to one or more constraints so that an output of the first objective function is minimized when the one or more constraints are satisfied; generating a path trajectory using the optimized first objective function to automatically control the autonomous driving vehicle according to the path trajectory; Controlling the autonomous driving vehicle to maneuver from the first position to the second position along the generated path trajectory; determining a second route from the second position of the autonomous vehicle to a third position within a second open space, the second position being a current position of the autonomous vehicle; determining a second objective function based on the second route, the second objective function comprising a second set of costs for maneuvering from the second location to the third location; determining an environmental condition of the second open space; as well as determining a second set of weights using the environmental conditions of the second open space, each weight in the second set of weights to be applied to a corresponding cost of the second objective function, at least one of the weights in the second set of weights being different from a corresponding weight in the first set of weights; Determining the first set of weights includes performing a table lookup in a table associating environmental conditions with weights representing the behavior of an expert driver with respect to the environmental conditions.

2. The method according to claim 1, wherein: The table is an inverse reinforcement learning model of the weights.

3. The method according to claim 1, further comprising: Determining the environmental condition of the first open space includes: Analyze images captured by one or more cameras of the autonomous vehicle to identify the environmental conditions.

4. The method according to claim 3, wherein: The environmental conditions include the position of one or more objects captured within the image relative to the current position of the autonomous vehicle.

5. The method according to claim 3, wherein: The environmental conditions include road features of the first open space captured within the image.

6. A non-transitory machine-readable medium having instructions stored therein, the instructions, when executed by a processor, causing the processor to perform operations comprising: determining a first route from a first location of the autonomous vehicle to a second location within a first open space, the first location being a current location of the autonomous vehicle; determining a first objective function based on the first route, the first objective function having a first set of costs for maneuvering the autonomous vehicle from the first location to the second location; determining a first set of weights based on one or more environmental conditions of the first open space, each weight in the first set of weights to be applied to a corresponding cost of the first objective function; Optimizing the first objective function according to one or more constraints so that an output of the first objective function is minimized when the one or more constraints are satisfied; generating a path trajectory using the optimized first objective function to automatically control the autonomous driving vehicle according to the path trajectory; Controlling the autonomous driving vehicle to maneuver from the first position to the second position along the generated path trajectory; determining a second route from the second position of the autonomous vehicle to a third position within a second open space, the second position being a current position of the autonomous vehicle; determining a second objective function based on the second route, the second objective function comprising a second set of costs for maneuvering from the second location to the third location; determining an environmental condition of the second open space; as well as determining a second set of weights using the environmental conditions of the second open space, each weight in the second set of weights to be applied to a corresponding cost of the second objective function, at least one of the weights in the second set of weights being different from a corresponding weight in the first set of weights; Determining the first set of weights includes performing a table lookup in a table associating environmental conditions with weights representing the behavior of an expert driver with respect to the environmental conditions.

7. The machine-readable medium of claim 6, wherein: The table is an inverse reinforcement learning model of the weights.

8. The machine-readable medium of claim 6, wherein: The operation further includes: determining the environmental condition of the first open space, including: Images captured by one or more cameras of the autonomous vehicle are analyzed to identify the environmental conditions.

9. Data processing system, including: processor; as well as a memory coupled to the processor to store instructions that, when executed by the processor, cause the processor to perform operations comprising: determining a first route from a first location of the autonomous vehicle to a second location within a first open space, the first location being a current location of the autonomous vehicle; determining a first objective function based on the first route, the first objective function having a first set of costs for maneuvering the autonomous vehicle from the first location to the second location; determining a first set of weights based on one or more environmental conditions of the first open space, each weight in the first set of weights to be applied to a corresponding cost of the first objective function; Optimizing the first objective function according to one or more constraints so that an output of the first objective function is minimized when the one or more constraints are satisfied; generating a path trajectory using the optimized first objective function to automatically control the autonomous driving vehicle according to the path trajectory; Controlling the autonomous driving vehicle to maneuver from the first position to the second position along the generated path trajectory; determining a second route from the second position of the autonomous vehicle to a third position within a second open space, the second position being a current position of the autonomous vehicle; determining a second objective function based on the second route, the second objective function comprising a second set of costs for maneuvering from the second location to the third location; determining an environmental condition of the second open space; and determining a second set of weights using the environmental conditions of the second open space, each weight in the second set of weights to be applied to a corresponding cost of the second objective function, at least one of the weights in the second set of weights being different from a corresponding weight in the first set of weights; Determining the first set of weights includes performing a table lookup in a table associating environmental conditions with weights representing the behavior of an expert driver with respect to the environmental conditions.

10. The data processing system according to claim 9, wherein: The table is an inverse reinforcement learning model of the weights.

11. The data processing system according to claim 9, wherein: The operation further includes: determining the environmental condition of the first open space, including: Images captured by one or more cameras of the autonomous vehicle are analyzed to identify the environmental conditions.

Images