Robotic Safety Trajectories for Adaptive Emergency Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing real-time software control systems for physical machines, such as industrial robots, often rely on predetermined safety trajectories that may be suboptimal or unsafe, particularly when unexpected situations arise, as they do not adapt to the current state of the execution environment, potentially leading to collisions or unsafe stops.
Innovation Solution
A system that generates new safety trajectories in real-time using recent sensor data, allowing for adaptive control to ensure safe operation by switching to an emergency control system when an emergency situation occurs, which can execute the most recent safety trajectory to prevent collisions or other hazards.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If predetermined safety trajectories are used in emergency situations, then the control system is simple and fast to execute, but the safety trajectory may be suboptimal or unsafe when unexpected situations arise
Solution Approach 1:
The system pre-generates multiple candidate safety trajectories covering various emergency scenarios before operation. When an emergency occurs, the system selects and executes the most appropriate pre-computed trajectory, avoiding the need for real-time optimization while ensuring safety across diverse situations.
Solution Approach 2:
The system dynamically selects from multiple pre-computed safety trajectories based on the actual emergency situation detected by sensors. The selection adapts to the current state, allowing the system to choose the most appropriate safety trajectory rather than executing a single fixed trajectory, thereby improving safety without requiring complex real-time computation.
2Adaptability or versatility
If real-time safety trajectory generation is implemented, then adaptability to current state improves, but computational time and processing requirements increase
Solution Approach 1:
Multiple candidate safety trajectories are pre-computed covering various emergency scenarios. During operation, the system only needs to select from these pre-computed options based on sensor feedback, rather than generating a new trajectory from scratch during the emergency response, thus maintaining adaptability while minimizing computational time.
Solution Approach 2:
The system pre-computes more safety trajectories than strictly necessary, covering a broader range of possible emergency scenarios. This excessive preparation ensures that an appropriate trajectory is always available for any unexpected situation, while the actual real-time computation is limited to simple selection rather than full trajectory generation.
3Ease of manufacture
If linear safety trajectories are used for emergency stops, then the control is simple to implement, but collisions may occur with other objects in the linear path
Solution Approach 1:
Multiple candidate safety trajectories including non-linear paths are pre-computed to account for objects in the execution environment. The system selects from these diverse options based on the current state, ensuring collision-free paths are available without requiring complex real-time path planning during emergencies.
Solution Approach 2:
The system dynamically selects from multiple pre-computed trajectories with different geometric characteristics (linear, curved, etc.) based on the detected environment. This allows the system to choose non-linear trajectories when objects are present in linear paths, reducing collision risk while maintaining implementation simplicity through pre-computation.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on computer storage media, for generating safety trajectories. One of the methods comprises causing, by a robotic control system, execution of a robotic control plan by a plurality of robotic components in a robotic execution environment; generating a safety trajectory at each of a plurality of time points during the execution of the robotic control plan, including: obtaining data identifying a current position of a particular robotic component of the plurality of robotic components; generating, using the obtained data, a safety trajectory for the particular robotic component; and providing the safety trajectory to an emergency control system; determining that an emergency condition has been met; and in response, transferring control of the particular robotic component from the robotic control system to the emergency control system, comprising causing, by the emergency control system, execution of the safety trajectory by the particular robotic component.