Autonomous Vehicle Task Allocation Using Search Tree Pruning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current task allocation methods for autonomous vehicles, such as genetic algorithms, mixed integer linear programming, and branch and bound, face challenges in finding optimal solutions within reasonable time due to high computational complexity and the difficulty in refining performance, especially in scenarios like the traveling salesman problem and search and rescue operations.
Innovation Solution
A task allocation system that determines candidate task sequences, generates a task allocation tree, and performs a branch and bound search to find a task allocation plan meeting predetermined criteria like lowest cost, incorporating a task sequencing module, allocation module, and execution module to control autonomous vehicles effectively.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If exhaustive search methods (branch and bound) are used to guarantee optimal solution, then solution optimality is improved, but processing time becomes unreasonably long
Solution Approach 1:
The patent segments the exhaustive search process by introducing admissible heuristics that divide the search space into manageable sections. Instead of uniformly exploring all possibilities, the heuristic functions guide the search to prioritize promising regions while still guaranteeing optimality by systematically checking bounds, thus reducing overall processing time while maintaining solution optimality.
Solution Approach 2:
The patent performs preliminary actions by computing admissible heuristic values for each node before actual path exploration. These pre-computed heuristic estimates allow the algorithm to prune entire branches of the search tree that cannot lead to optimal solutions, significantly reducing processing time while preserving the guarantee of finding the optimal path.
2Ease of operation
If simple algorithms are used for quick implementation, then ease of operation is improved, but solution quality and performance refinement become difficult
Solution Approach 1:
The patent introduces admissible heuristics as intermediary functions that bridge simple algorithmic structures with complex optimization requirements. These heuristics act as mediators that guide the search process without requiring complex algorithmic changes, maintaining ease of implementation while significantly improving solution quality through informed search direction.
Solution Approach 2:
The patent changes the parameter of search evaluation by incorporating heuristic cost estimates into the node selection process. This parameter change transforms a simple unguided search into an informed search that prioritizes promising paths, improving solution quality while maintaining the fundamental simplicity of the search framework.
3Reliability
If continuous optimization methods (mixed integer linear programming) are used, then mathematical rigor is improved, but discrete choices cannot always map back to discrete solutions
Solution Approach 1:
The patent inverts the conventional approach by working backward from the discrete solution space using admissible heuristics. Instead of relaxing discrete constraints to continuous variables and hoping for valid mappings, the heuristic-guided search operates directly in the discrete space, guaranteeing that all solutions remain valid discrete assignments while maintaining mathematical rigor through systematic exploration and pruning.
Data Source
AI summary
Methods and apparatus are provided for allocating tasks to be performed by one or more autonomous vehicles to achieve a mission objective. Generally, a task allocation system identifies a final task associated with a given mission objective, identifies predecessor tasks necessary to complete the final task, generates one or more candidate tasks sequences to accomplish the mission objective, generates a task allocation tree based on the candidate task sequences, and searches the task allocation tree to find a task allocation plan that meets a predetermined selection criteria (e.g., lowest cost). Based on the task allocation plan, the task allocation system determines a task execution plan and generates control data for controlling one or more autonomous vehicles to complete the task execution plan.