Behavior Engine for Remote Vehicle Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing remote vehicle behavior systems face scalability issues due to a lack of separation between action selection search processes and behavioral logic, leading to complex configurations and limited growth potential, particularly in managing 'big behaviors' and state machines.
Innovation Solution
A robotics system utilizing a behavior tree architecture with a control arbitration system that separates action selection from behavioral logic, employing a kinodynamic fixed depth motion planning algorithm to generate commands based on goals and constraints, and allowing dynamic resource management and behavior composition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a tree structure behavior system is used to manage remote vehicle behaviors, then behavioral logic can be organized hierarchically, but the system becomes prohibitively complex as it grows due to lack of separation between action selection search and state machine logic
Solution Approach 1:
The patent segments the behavior system into distinct modules: a behavior tree manager that handles high-level behavioral logic and a separate action selection search mechanism that handles low-level actuator command generation. This separation allows each component to be developed, maintained, and scaled independently, resolving the contradiction between system organization and complexity management.
Solution Approach 2:
The patent introduces an intermediary layer (the behavior tree manager) that sits between the high-level behavioral logic and the low-level action selection search. This intermediary translates behavioral goals into search constraints and coordinates the interaction between the two previously coupled systems, enabling independent evolution of each component while maintaining system coherence.
2Ease of operation
If big behaviors are statically configured as services and invoked ad-hoc via behavior system events, then flexibility in behavior invocation is achieved, but scalability problems arise due to hidden API and complex manager relationships
Solution Approach 1:
The patent extracts the action selection search functionality from the statically configured big behaviors and places it in a dedicated, reusable search mechanism. This extraction eliminates the hidden API problem by making the search interface explicit and standardized, while allowing big behaviors to be defined through clean, high-level interfaces that don't require knowledge of the underlying search complexity.
3Reliability
If each complex behavior creates its own state machine logic internally, then behavior-specific control is achieved, but large and complex monster behaviors result that limit system growth
Solution Approach 1:
The patent merges the state machine logic from individual behaviors into the behavior tree manager, which naturally provides hierarchical state management through the tree structure. Individual behaviors no longer need to create their own state machines; instead, they define their goals and constraints, while the behavior tree manager handles the coordination and state transitions, reducing implementation complexity while maintaining control accuracy.
Data Source
AI summary
A method of commanding a remote vehicle includes executing a command on a controller of the remote vehicle based on a kinodynamic fixed depth motion planning algorithm to use incremental feedback from evaluators to select a best feasible action. The method also includes determining servo commands corresponding to the best feasible action for one or more actuators of a drive system or a manipulation system of the remote vehicle and commanding the one or more actuators of the remote vehicle based on the servo commands. The best feasible action includes actions within a fixed time horizon of several seconds from a current time each time a feasible action is selected.