Robot Navigation Control Using Person Speed-Based Mode Switching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous transport robots face challenges in efficiently navigating around people due to unpredictable human movement, leading to difficulties in appropriate control and potential collisions.
Innovation Solution
A control system that uses machine learning to classify individuals into groups based on features extracted from camera images, estimating their moving speed and switching between high-load and low-load modes to adjust processing loads, camera settings, and control strategies, including switching control on mobile robots and putting cameras to sleep when not needed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the transport robot moves at low speed when a person is present, then safety is improved, but transport efficiency deteriorates
Solution Approach 1:
The robot dynamically adjusts its operating mode between high-load and low-load modes based on real-time detection of person presence and movement characteristics. This dynamic adaptation allows the robot to operate at high speed when safe (improving efficiency) and switch to low-speed cautious mode only when necessary (maintaining safety), rather than maintaining a fixed low speed that would always reduce efficiency.
Solution Approach 2:
The system changes operational parameters (speed, processing load, camera settings) based on the detected state of the environment. When a person is detected, the system evaluates their movement speed and direction, then adjusts robot speed and camera frame rate accordingly. This parameter adaptation resolves the contradiction by optimizing both safety and efficiency under different conditions.
2Reliability
If high-load processing is performed to accurately detect and respond to people, then safety is improved, but energy consumption and processing load increase
Solution Approach 1:
The system applies partial processing action by adjusting camera frame rates and processing loads based on situation. When no person is present or the person is moving quickly away, the system reduces processing load (lower frame rates, simplified detection). When a person is present and moving slowly (potentially dangerous), the system increases processing load (higher frame rates, full detection algorithms). This partial action approach maintains detection accuracy when needed while reducing energy consumption during normal operation.
Solution Approach 2:
The processing load dynamically adapts to the situation. The system switches between high-load mode (full camera resolution, high frame rate, comprehensive person detection algorithms) and low-load mode (reduced resolution, lower frame rate, simplified detection). This dynamic adjustment ensures detection accuracy is maintained when safety requires it, while minimizing energy consumption during routine operation.
3Reliability
If the robot continuously monitors and adjusts control based on person detection, then safety is improved, but system complexity increases
Solution Approach 1:
The control system is segmented into distinct functional modules: camera/image processing module, person detection module, movement analysis module, and control decision module. Each module handles a specific aspect of the safety control process. This segmentation makes the complex system more manageable and maintainable while enabling continuous monitoring and adjustment based on person detection.
Solution Approach 2:
The system introduces an intermediary processing layer that bridges raw camera data and robot control commands. This intermediary layer includes image processing, person detection algorithms, and movement analysis that translate complex visual data into simple control decisions (speed adjustment, direction changes). This intermediary structure manages system complexity by breaking down the control process into manageable stages.
Data Source
AI summary
A control system comprises one or more processors. The one or more processors are configured to extract a feature of a person in an image captured by a camera, classify the person into a preset first group or a preset second group based on the feature, estimate a moving speed of the person belonging to the second group, and switch, based on the moving speed, a mode between a high-load mode for performing a high-load process and a low-load mode for performing a process with a load lower than a load in the high-load mode.