Path generation based on predicted actions
Provided are methods and systems for semantic behavior filtering for prediction improvement. A method for operating an autonomous vehicle is provided. The method includes obtaining, by at least one processor, semantic image data associated with an environment in which an autonomous vehicle is operating. The method includes determining, by the at least one processor a set of agents in the environment based on the semantic image data. The method includes determining, by the at least one processor, a set of secondary agents from the set of agents based on a relative location of a respective secondary agent to a respective object and a set of object semantic behavior data associated with the respective object. The method includes determining, from the set of agents, a set of primary agents other than secondary agents. The method includes generating a path for the autonomous vehicle based on the set of primary agents.
1 . A method for operating an autonomous vehicle, the method comprising:
obtaining, by at least one processor, semantic image data associated with an environment in which an autonomous vehicle is operating;
determining, by the at least one processor, a set of agents in the environment based on the semantic image data comprising feature embeddings corresponding to object attributes, wherein the set of agents comprise moveable objects;
determining, by the at least one processor, a set of secondary agents from the set of agents based on the feature embeddings and a relative location of a respective secondary agent to a respective object separate from the autonomous vehicle and a set of object semantic behavior data associated with the respective object, wherein the set of object semantic behavior data comprises logic-based rules and exceptions for safe automotive operation;
determining, from the set of agents, a set of primary agents based on the feature embeddings, wherein the set of primary agents are distinguished from the set of secondary agents based on a likelihood of interaction of agents within the set of primary agents with the autonomous vehicle satisfying a probability threshold and a likelihood of interaction of agents within the set of secondary agents with the autonomous vehicle not satisfying the probability threshold as determined based on the feature embeddings corresponding to the object attributes; and
generating a path for the autonomous vehicle based on the set of primary agents by at least determining trajectories of the set of primary agents while removing a subset of neural network trajectory predictions by excluding determining trajectories of the set of secondary agents, wherein generating the path comprises:
determining primary actions and secondary actions for a primary agent of the set of primary agents, wherein the primary actions and the secondary actions correspond to movements, and wherein primary actions satisfy an action probability threshold of occurring and secondary actions do not satisfy the action probability threshold of occurring;
using the primary actions of the primary agent to generate the path; and
excluding the secondary actions of the primary agent to generate the path.
2 . The method of claim 1 , further comprising causing the autonomous vehicle to operate along the path for the autonomous vehicle.
3 . The method of claim 1 , wherein the set of agents comprise objects configured to move.
4 . The method of claim 1 , wherein the set of secondary agents comprises agents associated with less than a threshold probability of interacting with the autonomous vehicle.
5 . The method of claim 1 , wherein the set of secondary agents comprises agents associated with greater than a threshold probability of interacting with the autonomous vehicle.
6 . The method of claim 1 , wherein the set of secondary agents comprises an agent disposed on an opposite side of a traffic signal from the autonomous vehicle.
7 . The method of claim 1 , wherein agents of the set of secondary agents are vehicles approaching a traffic signal opposite the autonomous vehicle.
8 . A system, comprising:
at least one processor, and
at least one non-transitory storage media storing instructions that, when executed by the at least one processor, cause the at least one processor to:
obtain semantic image data associated with an environment in which an autonomous vehicle is operating, wherein the semantic image data comprises feature embeddings corresponding to object attributes;
determine a set of agents in the environment based on the feature embeddings of the semantic image data, wherein the set of agents comprise moveable objects;
determine a set of secondary agents from the set of agents based on the feature embeddings and a relative location of a respective secondary agent to a respective object separate from the autonomous vehicle and a set of object semantic behavior data associated with the respective object, wherein the set of object semantic behavior data comprises logic-based rules and exceptions for safe automotive operation;
determine, from the set of agents, a set of primary agents based on the feature embeddings, wherein the set of primary agents are distinguished from the set of secondary agents based on a likelihood of interaction of agents within the set of primary agents with the autonomous vehicle satisfying a probability threshold and a likelihood of interaction of agents within the set of secondary agents with the autonomous vehicle not satisfying the probability threshold as determined based on the feature embeddings corresponding to the object attributes; and
generate a path for the autonomous vehicle based on the set of primary agents by at least determining trajectories of the set of primary agents while removing a subset of neural network trajectory predictions by excluding determining trajectories of the set of secondary agents, wherein generating the path comprises:
determining primary actions and secondary actions for a primary agent of the set of primary agents, wherein the primary actions and the secondary actions correspond to movements, and wherein primary actions satisfy an action probability threshold of occurring and secondary actions do not satisfy the action probability threshold of occurring;
using the primary actions of the primary agent to generate the path; and
excluding the secondary actions of the primary agent to generate the path.
9 . The system of claim 8 , further comprising causing the autonomous vehicle to operate along the path for the autonomous vehicle.
10 . The system of claim 8 , wherein the set of agents comprise objects configured to move.
11 . The system of claim 8 , wherein the set of secondary agents comprises agents associated with less than a threshold probability of interacting with the autonomous vehicle.
12 . The system of claim 8 , wherein the set of secondary agents comprises agents associated with greater than a threshold probability of interacting with the autonomous vehicle.
13 . The system of claim 8 , wherein the set of secondary agents comprises an agent disposed on an opposite side of a traffic signal from the autonomous vehicle.
14 . The system of claim 8 , wherein agents of the set of secondary agents are vehicles approaching a traffic signal opposite the autonomous vehicle.
15 . At least one non-transitory storage media storing instructions that, when executed by at least one processor, cause the at least one processor to:
obtain semantic image data associated with an environment in which an autonomous vehicle is operating, wherein the semantic image data comprises feature embeddings corresponding to object attributes;
determine a set of agents in the environment based on the feature embeddings of the semantic image data, wherein the set of agents comprise moveable objects;
determine a set of secondary agents from the set of agents based on the feature embeddings and a relative location of a respective secondary agent to a respective object separate from the autonomous vehicle and a set of object semantic behavior data associated with the respective object, wherein the set of object semantic behavior data comprises logic-based rules and exceptions for safe automotive operation;
determine, from the set of agents, a set of primary agents based on the feature embeddings, wherein the set of primary agents are distinguished from the set of secondary agents based on a likelihood of interaction of agents within the set of primary agents with the autonomous vehicle satisfying a probability threshold and a likelihood of interaction of agents within the set of secondary agents with the autonomous vehicle not satisfying the probability threshold as determined based on the feature embeddings corresponding to the object attributes; and
generate a path for the autonomous vehicle based on the set of primary agents by at least determining trajectories of the set of primary agents while removing a subset of neural network trajectory predictions by excluding determining trajectories of the set of secondary agents, wherein generating the path comprises:
determining primary actions and secondary actions for a primary agent of the set of primary agents, wherein the primary actions and the secondary actions correspond to movements, and wherein primary actions satisfy an action probability threshold of occurring and secondary actions do not satisfy the action probability threshold of occurring;
using the primary actions of the primary agent to generate the path; and
excluding the secondary actions of the primary agent to generate the path.
16 . The at least one non-transitory storage media of claim 15 , further comprising causing the autonomous vehicle to operate along the path for the autonomous vehicle.
17 . The at least one non-transitory storage media of claim 15 , wherein the set of agents comprise objects configured to move.
18 . The at least one non-transitory storage media of claim 15 , wherein the set of secondary agents comprises agents associated with less than a threshold probability of interacting with the autonomous vehicle.
19 . The at least one non-transitory storage media of claim 15 , wherein the set of secondary agents comprises agents associated with greater than a threshold probability of interacting with the autonomous vehicle.
20 . The at least one non-transitory storage media of claim 15 , wherein the set of secondary agents comprises an agent disposed on an opposite side of a traffic signal from the autonomous vehicle.