Secondary fallback software stack for autonomous vehicle
Systems and methods for a vehicle to use a fallback control system orthogonal to a primary control system are provided. A method includes collecting, by a fallback control system of a vehicle, sensor data generated by one or more sensors of the vehicle; determining, by the fallback control system using a first process different from a second process of a primary control system of the vehicle, an alternative action for the vehicle based on the sensor data; and sending, by the fallback control system based on the alternative action, an instruction for modifying an action associated with the primary control system upon a faulty condition associated with the primary control system.
1. A method comprising:
determining, by a primary planner of an autonomous vehicle, a first action for the autonomous vehicle using a first process;
determining, by a fallback planner of the autonomous vehicle independent of the primary planner, an alternative action for the autonomous vehicle using a second process nonoverlapping with the first process, wherein the alternative action includes at least stopping the autonomous vehicle;
sending, by a multiplexer of a common control system of the autonomous vehicle, an indication of the alternative action for actuators to use to cause the autonomous vehicle to navigate in response to a faulty condition associated with the primary planner, wherein the faulty condition is the primary planner failing to predict a trajectory for other traffic participants in an environment of the autonomous vehicle; and
computing, by the fallback planner, a collision-free path for the autonomous vehicle responsive to a determination of an area at which the autonomous vehicle stopped after performing the alternative action fails to satisfy one or more area criteria.
2. The method of claim 1 , wherein the computing the collision-free path for the vehicle comprises:
determining candidate exit roads, each including one or more exit lanes;
filtering the exit lanes of the candidate exit roads based on one or more exit criteria; and
selecting a first exit lane from the filtered exit lanes based at least in part on a distance between a location of the vehicle and the first exit lane.
3. The method of claim 2 , wherein the computing the collision-free path for the vehicle further comprises:
determining whether a path from the location of the vehicle to the first exit lane is collision-free.
4. The method of claim 1 , wherein:
the first process used by the primary planner for determining the first action is based on a non-convex solver; and
the second process used by the fallback planner for determining the alternative action is based on a free space planner.
5. The method of claim 1 , wherein the computing the collision-free path is further based on free space planning, and wherein the free space planning is based on:
an occupancy grid map including an indication of one or more free spaces; and
motion information associated with one or more tracked objects in the occupancy grid map.
6. The method of claim 1 , further comprising:
determining the area at which the vehicle stopped fails to satisfy the one or more area criteria based on the area being associated with at least one of an intersection, an oncoming traffic lane, or an emergency driveway.
7. The method of claim 1 , further comprising:
determining the area at which the vehicle stopped after performing the alternative action fails to satisfy the one or more area criteria,
wherein the one or more area criteria is associated with an imminent collision between the vehicle and an object in the area.
8. The method of claim 1 , wherein the determining the alternative action by the fallback planner is performed concurrently with the determining the first action by the primary planner.