Non-blocking boundary for autonomous vehicle planning
According to some embodiments, a system generates a driving trajectory from a starting point to a destination point for an ADV. In one embodiment, the system calculates a first trajectory based on a map and a route information. The system generates a path profile based on the first trajectory, traffic rules, and blocking obstacles perceived by the ADV. The system determines non-blocking obstacles perceived by the ADV. The system generates a speed profile of the ADV for the path profile based on the non-blocking obstacles to identify a speed of the ADV to avoid the blocking obstacles in view of the non-blocking obstacles. The system generates a second trajectory based on the path profile and the speed profile to control the ADV autonomously according to the second trajectory.
1. A computer-implemented method to generate a driving trajectory for an autonomous driving vehicle (ADV), the method comprising:
calculating a first trajectory based on a map and a route information;
generating a path profile based on the first trajectory, traffic rules, and blocking obstacles perceived by the ADV, wherein the path profile and a station-time (ST) boundary of each of the blocking obstacles are modeled on an ST graph, wherein the ST boundary represents a set of points describing the blocking obstacle is to block the ADV at a particular position at a point of time;
determining non-blocking obstacles perceived by the ADV, including projecting predetermined regions outwardly from a left and a right side of the first trajectory and determining if obstacles perceived by the ADV will overlap the predetermined regions;
modeling an ST boundary of each of the determined non-blocking obstacles on the ST graph;
generating, based on the ST graph, a speed profile of the ADV for the path profile based on the ST boundaries of the non-blocking obstacles, and the ST boundaries of the blocking obstacles to identify a speed of the ADV to avoid the blocking obstacles in view of the non-blocking obstacles; and
generating a second trajectory based on the path profile and the speed profile to control the ADV autonomously according to the second trajectory.
2. The computer-implemented method of claim 1 , wherein generating the speed profile further comprises performing a dynamic optimization on a speed of the ADV based on the path profile and in view of the non-blocking obstacles.
3. The computer-implemented method of claim 1 , wherein a non-blocking obstacle of the non-blocking obstacles corresponds to a decision not to bypass a neighboring vehicle.
4. The computer-implemented method of claim 1 , wherein a non-blocking obstacle of the non-blocking obstacles corresponds to a decision to keep clear from a blind zone region of a neighboring vehicle.
5. The computer-implemented method of claim 1 , wherein a non-blocking obstacle of the non-blocking obstacles corresponds to a decision to keep clear from a particular road region.
6. The computer-implemented method of claim 1 , wherein performing a dynamic optimization comprises optimizing a cost function to generate the ST graph for the speed profile.
7. The computer-implemented method of claim 6 , wherein the cost function includes a weight factor for each non-blocking and blocking obstacles, wherein a higher weight is assigned to a blocking obstacle than a non-blocking obstacle.
8. The computer-implemented method of claim 6 , wherein a weight factor assigned to a non-blocking obstacle corresponding to a neighboring vehicle is dependent on a size of the neighboring vehicle perceived by the ADV.
9. The computer-implemented method of claim 6 , wherein the cost function comprises a velocity cost.
10. A non-transitory machine-readable medium having instructions stored therein, which when executed by a processor, cause the processor to perform operations, the operations comprising:
calculating a first trajectory based on a map and a route information;
generating a path profile based on the first trajectory, traffic rules, and blocking obstacles perceived by the ADV, wherein the path profile and a station-time (ST) boundary of each of the blocking obstacles are modeled on an ST graph, wherein the ST boundary represents a set of points describing the blocking obstacle is to block the ADV at a particular position at a point of time;
determining non-blocking obstacles perceived by the ADV, including projecting predetermined regions outwardly from a left and a right side of the first trajectory and determining if obstacles perceived by the ADV will overlap the predetermined regions;
modeling an ST boundary of each of determined non-blocking obstacles on the ST graph;
generating, based on the ST graph, a speed profile of the ADV for the path profile based on the ST boundaries of the non-blocking obstacles, and the ST boundaries of the blocking obstacles to identify a speed of the ADV to avoid the blocking obstacles in view of the non-blocking obstacles; and
generating a second trajectory based on the path profile and the speed profile to control the ADV autonomously according to the second trajectory.
11. The non-transitory machine-readable medium of claim 10 , wherein generating the speed profile further comprises performing a dynamic optimization on a speed of the ADV based on the path profile and in view of the non-blocking obstacles.
12. The non-transitory machine-readable medium of claim 10 , wherein a non-blocking obstacle of the non-blocking obstacles corresponds to a decision not to bypass a neighboring vehicle.
13. The non-transitory machine-readable medium of claim 10 , wherein a non-blocking obstacle of the non-blocking obstacles corresponds to a decision to keep clear from a blind zone region of a neighboring vehicle.
14. The non-transitory machine-readable medium of claim 10 , wherein a non-blocking obstacle of the non-blocking obstacles corresponds to a decision to keep clear from a particular road region.
15. The non-transitory machine-readable medium of claim 10 , wherein performing a dynamic optimization comprises optimizing a cost function to generate the ST graph for the speed profile.
16. The non-transitory machine-readable medium of claim 15 , wherein the cost function includes a weight factor for each non-blocking and blocking obstacles, wherein a higher weight is assigned to a blocking obstacle than a non-blocking obstacle.
17. The non-transitory machine-readable medium of claim 14 , wherein a weight factor assigned to a non-blocking obstacle corresponding to a neighboring vehicle is dependent on a size of the neighboring vehicle perceived by the ADV.
18. The non-transitory machine-readable medium of claim 15 , wherein the cost function comprises a velocity cost.
19. A data processing system, comprising:
a processor; and
a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations, the operations including calculating a first trajectory based on a map and a route information;
generating a path profile based on the first trajectory, traffic rules, and blocking obstacles perceived by the ADV, wherein the path profile and a station-time (ST) boundary of each of the blocking obstacles are modeled on an ST graph, wherein the ST boundary represents a set of points describing the blocking obstacle is to block the ADV at a particular position at a point of time;
determining non-blocking obstacles perceived by the ADV, including projecting predetermined regions outwardly from a left and a right side of the first trajectory and determining if obstacles perceived by the ADV will overlap the predetermined regions;
modeling an ST boundary of each of the determined non-blocking obstacles on the ST graph;
generating, based on the ST graph, a speed profile of the ADV for the path profile based on the ST boundaries of the non-blocking obstacles, and the ST boundaries of the blocking obstacles to identify a speed of the ADV to avoid the blocking obstacles in view of the non-blocking obstacles; and
generating a second trajectory based on the path profile and the speed profile to control the ADV autonomously according to the second trajectory.