Image Processing Control for Autonomous Driving Stalemates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing autonomous driving systems often encounter stalemates with targets, such as pedestrians or flagmen, due to inability to recognize unclear gestures or intentions, leading to obstructed traffic and reduced convenience.
Innovation Solution
An image processing apparatus mounted on a movable object that senses the state of a target in captured imagery, determines a first action based on the target's state, and subsequently determines a second action based on the target's state after the first action, with the second action being output at a time dependent on the target's motion speed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the autonomous driving system determines actions based on target state sensing, then the system can respond to target movements, but stalemates occur when targets make unclear gestures or intentions
Solution Approach 1:
The system dynamically adjusts the action determination timing based on the target's motion speed. When the target moves quickly, the system determines actions more frequently to capture rapid changes in gesture or intention. This dynamic adaptation allows the system to handle unclear gestures more effectively by increasing observation frequency, thereby resolving stalemates while maintaining reliable action determination.
Solution Approach 2:
The system implements a feedback mechanism where the determined action is output to the movable object, which then executes the action and its effect is observed in subsequent target state sensing. This closed-loop feedback allows the system to learn from the outcomes of its actions and adjust future determinations, improving reliability in handling ambiguous target intentions and preventing traffic stalemates.
2Extent of automation
If the system outputs action information to the movable object, then the movable object can perform actions, but traffic flow may be obstructed when stalemates occur
Solution Approach 1:
The system dynamically adjusts the timing of action information output based on real-time target motion speed. By increasing the frequency of action determination and output when targets move quickly or exhibit unclear gestures, the system maintains high automation while preventing stalemates that would obstruct traffic flow, thus preserving both autonomous execution capability and traffic efficiency.
Solution Approach 2:
The system changes the parameter of action determination timing based on target motion speed. When motion speed increases or gestures become unclear, the system adjusts the determination frequency parameter upward, allowing more frequent action outputs. This parameter adaptation enables the autonomous system to maintain productivity by preventing stalemates while preserving the extent of automation.
3Measurement precision
If the system determines actions based on captured imagery, then the system can sense target state, but the response time may not be appropriate for fast-moving targets
Solution Approach 1:
The system dynamically adjusts the action determination timing based on target motion speed. When targets move quickly, the system reduces the time interval between successive action determinations, allowing it to capture rapid changes in target state with appropriate precision while minimizing response time loss. This dynamic timing adjustment resolves the contradiction between measurement precision and response time.
Solution Approach 2:
The system changes the timing parameter of action determination based on target motion speed. By adjusting the determination frequency parameter upward for fast-moving targets, the system maintains accurate target state sensing while reducing the time loss between sensing and response, effectively resolving the contradiction between precision and response time.
Data Source
AI summary
An image processing apparatus 10 includes an output unit 14 and a processor 16. The output unit 14 outputs to a movable object 12 information indicating an action to be performed by the movable object 12. The processor 16 determines a first action of the movable object 12 based on a state of a target, the state being sensed in imagery of a region around the movable object 12. The processor 16 causes the output unit 14 to output information indicating the first action. The processor 16 determines a second action of the movable object 12 based on a state of the target, the state being sensed in imagery of the region around the movable object 12, the imagery being captured after the first action of the movable object 12. The processor 16 causes the output unit 14 to output information indicating the second action at a time depending on a speed of a predetermined motion of the target. The predetermined motion of the target is a motion that the target makes before the first action is determined or after the first action of the movable object 12.