Methods and systems for performing inter-trajectory re-linearization about an evolving reference path for an autonomous vehicle
Disclosed herein are systems, methods, and computer program products for generating a reference path for an autonomous vehicle. The methods comprising: receiving, by a computing device, a reference trajectory for the autonomous vehicle; generating, by the computing device, a first alternative trajectory for the autonomous vehicle based on an objective identified by the reference trajectory; comparing, by the computing device, the first alternative trajectory to the reference trajectory or a previously generated alternative trajectory; selecting, by the computing device, whether to generate a second alternative trajectory based on a result of said comparing; and generating, by the computing device, the second alternative trajectory for the autonomous vehicle in response to said selecting.
1. A method for generating a trajectory for an autonomous vehicle, comprising:
performing, by a computing device, a process comprising:
projecting an objective identified by a reference trajectory from a Cartesian reference frame to a curvilinear coordinate frame as a constraint indicating where the autonomous vehicle should be laterally in order to satisfy the objective;
generating an alternative trajectory having a refined lateral path for the autonomous vehicle based on the objective projected in the curvilinear coordinate; and
mapping the alternative trajectory having the refined lateral path from the curvilinear coordinate frame to the Cartesian reference frame;
comparing, by the computing device, the first alternative trajectory to the reference trajectory or a previously generated alternative trajectory to obtain at least one discrepancy measure;
selecting, by the computing device, whether to repeat the process based on the at least one discrepancy measure; and
repeating, by the computing device, the process using the alternative trajectory, in response to said selecting.
2. The method according to claim 1 , wherein said generating the alternative trajectory comprises decoupling a lateral component of the objective and a longitudinal component of the objective.
3. The method according to claim 2 , wherein said generating the alternative trajectory further comprises linearizing the lateral component about the reference trajectory to obtain a lateral problem.
4. The method according to claim 3 , wherein said generating the alternative trajectory further comprises solving the lateral problem to obtain a linearized lateral component of the objective.
5. The method according to claim 4 , wherein said generating the alternative trajectory comprises fusing the linearized lateral component with a longitudinal component of the objective to create a fused solution, as the refined lateral path.
6. The method according to claim 1 , wherein said comparing comprises projecting the alternative trajectory onto the reference trajectory or the previously generated alternative trajectory to obtain a convergence measure.
7. The method according to claim 6 , wherein the convergence measure is a function of a sequence of lateral discrepancy measures, as the discrepancy measure.
8. The method according to claim 1 , wherein the reference trajectory and, the alternative trajectory are different from each other.
9. A system, comprising:
a processor;
a non-transitory computer-readable storage medium comprising programming instructions that are configured to cause the processor to implement a method for generating a trajectory for an autonomous vehicle, wherein the programming instructions comprise instructions to:
perform a process comprising:
project an objective identified by a reference trajectory from a Cartesian reference frame to a curvilinear coordinate frame as a constraint indicating where the autonomous vehicle should be laterally in order to satisfy the objective;
generate an alternative trajectory having a refined lateral path for the autonomous vehicle based on the objective projected in the curvilinear coordinate; and
map the alternative trajectory having the refined lateral path from the curvilinear coordinate frame to the Cartesian reference frame;
compare the alternative trajectory to the reference trajectory or a previously generated alternative trajectory to obtain at least one discrepancy measure;
select whether to repeat the process based on the at least one discrepancy measure; and
repeat the process using the alternative trajectory, the second alternative trajectory for the autonomous vehicle in response to said selecting.
10. The system according to claim 9 , wherein the alternative trajectory is generated by further decoupling a lateral component of the objective and a longitudinal component of the objective.
11. The system according to claim 10 , wherein the alternative trajectory is generated by further linearizing the lateral component about the reference trajectory to obtain a lateral problem.
12. The system according to claim 11 , wherein the alternative trajectory is generated by further solving the lateral problem to obtain a linearized lateral component of the objective.
13. The system according to claim 12 , wherein the alternative trajectory is generated by further fusing the linearized lateral component with a longitudinal component of the objective to create a fused solution.
14. The system according to claim 9 , wherein the alternative trajectory is compared to the reference trajectory or a previously generated alternative trajectory by projecting the alternative trajectory onto the reference trajectory or the previously generated alternative trajectory to obtain a convergence measure.
15. The system according to claim 14 , wherein the convergence measure is a function of a sequence of lateral discrepancy measures, as the discrepancy measure.
16. A non-transitory computer-readable medium that stores instructions that are configured to, when executed by at least one computing device, cause the at least one computing device to perform operations comprising:
performing, by a computing device, a process comprising:
projecting an objective identified by a reference trajectory of an autonomous vehicle from a Cartesian reference frame to a curvilinear coordinate frame as a constraint indicating where the autonomous vehicle should be laterally in order to satisfy the objective;
generating an alternative trajectory having a refined lateral path for the autonomous vehicle based on an objective identified by the reference trajectory; and
mapping the alternative trajectory having the refined lateral path from the curvilinear coordinate frame to the Cartesian reference frame;
comparing the alternative trajectory to the reference trajectory or a previously generated alternative trajectory to obtain at least one discrepancy measure;
selecting whether to repeat the process based on the at least one discrepancy measure; and
repeating the process in response to said selecting.