Deadlock Detection for Mobile Vehicle Path Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for creating traveling plans for mobile vehicles in warehouses or factories often result in deadlocks due to complex path configurations and large numbers of vehicles, where not all states can be searched within a realistic time frame, leading to suboptimal or deadlock-causing plans.
Innovation Solution
A deadlock detection device that uses a combining calculator and a deadlock determiner to analyze the traveling path configuration graph and mobile vehicle states, identifying potential deadlocks by moving vehicles within tree structure parts and combining them to determine if all vehicles can reach intersections or terminal points, thereby avoiding exhaustive search and reducing simulation time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If simulation time period or number of times is increased to search more states, then optimal traveling plan can be found, but calculation time becomes excessively long and cannot be completed in realistically allowed time period
Solution Approach 1:
The patent segments the traveling path configuration graph into tree structure parts and non-tree structure parts. By focusing simulation only on tree structure parts (which have unique paths and cannot cause deadlocks), the search space is dramatically reduced while maintaining accuracy for deadlock detection in the non-tree parts.
Solution Approach 2:
The patent extracts and identifies tree structure parts from the overall traveling path configuration graph. These tree structure parts are then excluded from exhaustive simulation, as they cannot cause deadlocks. This extraction allows the system to focus computational resources only on the critical non-tree structure parts where deadlocks may occur.
2Loss of time
If upper limit of simulation time period or number of times is provided to reduce calculation time, then calculation can be completed in realistic time, but sufficient search process cannot be made leading to deadlock-causing plans
Solution Approach 1:
The patent divides the path configuration into tree structure parts (safe, no deadlock possible) and non-tree structure parts (potential deadlock zones). This segmentation allows unlimited simulation time allocation to the critical non-tree parts while minimizing or eliminating simulation in tree parts, ensuring deadlock-free guarantees without excessive calculation time.
Solution Approach 2:
The patent changes the simulation approach by identifying and excluding tree structure parts from the simulation scope. This parameter change (from simulating all paths to simulating only non-tree parts) maintains sufficient search depth for deadlock detection while reducing total calculation time to realistic levels.
3Reliability
If traditional deadlock prevention methods (rules, blocked sections, one-way paths) are applied, then deadlock occurrence can be prevented, but traveling plans become inefficient and cannot be optimal
Solution Approach 1:
The patent introduces tree structure identification as an intermediary mechanism between path configuration and deadlock detection. By using graph theory concepts (tree structures) as a mediator, the system can automatically identify safe zones without imposing restrictive rules, allowing flexible and efficient traveling plans while maintaining deadlock prevention.
4Measurement precision
If exhaustive state search is performed to ensure optimal traveling plan, then all possible states can be evaluated, but calculation time becomes prohibitively long for large numbers of vehicles and complex path configurations
Solution Approach 1:
The patent segments the state search space by identifying tree structure parts of the traveling path configuration. States involving only tree structure parts are recognized as safe and require no exhaustive search. This segmentation maintains measurement precision for deadlock detection by focusing exhaustive search only on non-tree structure parts where deadlocks can actually occur.
Solution Approach 2:
The patent extracts tree structure parts from the overall path configuration and removes them from the exhaustive search process. This extraction maintains completeness of state search for critical areas (non-tree parts) while eliminating redundant computation in safe areas (tree parts), reducing calculation time from prohibitively long to manageable levels.
Data Source
AI summary
According to one embodiment, a deadlock detection device includes a combining calculator and a deadlock determiner. The combining calculator performs selecting a mobile vehicle or combined mobile vehicles from among mobile vehicles, based on a traveling path configuration graph and first state information, going forward the selected mobile vehicle to go forward on traveling path configuration graph and combining the selected mobile vehicle to another mobile vehicle or another combined mobile vehicles at a back of the other mobile vehicle or the other combined mobile vehicles, iterating a process of the selecting, the going and combining. The deadlock determiner determines that a deadlock occurs if not all the mobile vehicles have been combined by the combining calculator, and determines that no deadlock occurs if all the mobile vehicles have been combined.