Robot Navigation Control Using Predicted Human Occupation Areas
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing mobile robot navigation systems fail to adequately consider human motion and contextual information, leading to increased collision risks and discomfort when interacting with humans in dynamic environments.
Innovation Solution
A method that integrates predicted geometric and contextual information, such as body pose and velocity, to generate control parameters that ensure a safe distance from humans, using model predictive control to navigate the robot device.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the robot uses basic geometric human motion prediction for navigation, then the navigation system is simple and fast, but the collision risk increases and human safety is compromised
Solution Approach 1:
The system performs preliminary prediction of human motion trajectories and occupation areas before the robot makes navigation decisions. By predicting where humans will be in future time periods and calculating their occupation areas in advance, the robot can plan safer paths proactively rather than reactively, reducing collision risk while maintaining systematic control
Solution Approach 2:
The patent extends the navigation consideration from simple 2D spatial paths to 3D occupation volumes by incorporating human body pose and velocity information. This dimensional expansion allows the robot to account for the spatial extent and dynamic boundaries of human bodies, not just their projected positions, thereby improving safety without excessive complexity
2Reliability
If the robot maintains large safety distances from humans, then collision risk is reduced, but the robot's mobility and productivity decrease
Solution Approach 1:
Instead of applying a uniform safety distance in all directions, the system calculates direction-specific occupation areas based on predicted human motion trajectories. The safety margin is locally adapted to each direction of potential human movement, allowing the robot to navigate closer to humans when their predicted paths indicate lower risk while maintaining larger margins in directions where humans are likely to move, thus balancing safety and mobility
Solution Approach 2:
The safety distance is made dynamic by continuously updating human occupation areas based on real-time velocity and pose predictions. As humans change their motion states, the occupation areas and corresponding safety margins are adjusted dynamically, allowing the robot to optimize its path in real-time rather than following fixed conservative distance rules, thereby improving both safety and productivity
3Productivity
If the robot navigates faster to improve productivity, then efficiency increases, but the discomfort of humans and collision risk increase
Solution Approach 1:
The system incorporates feedback from predicted human velocity and activity levels to dynamically adjust robot navigation speed. When humans are detected with higher velocities or in active states, the robot reduces its speed to match the social context, minimizing discomfort and collision risk. This feedback loop allows the robot to optimize speed based on real-time human behavior rather than maintaining constant high speed
4Reliability
If the robot uses detailed contextual human information for control, then navigation smoothness and safety improve, but computational requirements and processing time increase
Solution Approach 1:
The control system is segmented into modular components: human detection, pose estimation, velocity calculation, occupation area prediction, and trajectory planning. Each module processes specific aspects of human information independently and passes results to the next stage. This segmentation allows parallel processing and optimization of individual modules, reducing overall computation time while maintaining comprehensive safety analysis
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method (200) for controlling a robot device (102) is disclosed which includes: determining, using sensor data (302) representing a surrounding of the robot device, whether there are one or more humans (106) in the surrounding of the robot device; determining for each human of the one or more humans: a body pose (308) and a velocity (310) of the human, a predicted motion (306) of the human for a future time period, and, using the body pose (308), the velocity (310), and the predicted motion (306), an occupation area (320) occupied by the human in the future time period; generating control parameters (322) for controlling the robot device such that a predefined distance of the robot device (102) to the occupation area (320) of each of the one or more humans is ensured in the future time period; and controlling the robot device in accordance with the control parameters.