Decentralized Control Architecture for Humanoid Robot Limbs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current control-command architectures for humanoid robots with articulated limbs face limitations due to a single central computation unit managing all motors and sensors, leading to communication bottlenecks and saturation, which restricts the complexity and versatility required for advanced human-like behaviors and mission adaptability.
Innovation Solution
A decentralized control-command architecture with at least three levels: a microcontroller-based electronic card level for sensor/actuator control, a translation and transmission level, and a higher-level function generation including artificial intelligence, where control functions are distributed between two computers, allowing for secure and optimized communication and enabling interchangeable components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If a single central computation unit is used to control all motors and sensors, then the control architecture is simple, but communication bottlenecks and saturation occur, limiting robot complexity and versatility
Solution Approach 1:
The control architecture is segmented into three hierarchical levels: electronic cards controlling individual motor units, a first computer managing communication and coordination, and a second computer handling high-level functions. This segmentation distributes the computational load and eliminates communication bottlenecks while enabling greater robot complexity and versatility.
2Adaptability or versatility
If a decentralized architecture with a mother card and controller card for each motor unit is used, then communication capacity increases, but the processor becomes saturated managing communications
Solution Approach 1:
The decentralized architecture is further segmented by introducing a first computer that handles communication management and coordination, separating this task from the second computer that focuses on high-level functions. This segmentation prevents processor saturation while maintaining high communication capacity.
Solution Approach 2:
The first computer acts as an intermediary between the electronic cards and the second computer, managing communication protocols, coordinating data flow, and handling lower-level control tasks. This intermediary role offloads communication management from the second computer, preventing processor saturation while maintaining efficient communication across the system.
3Productivity
If control functions are distributed between two computers, then computation power for artificial intelligence is freed up, but system complexity increases
Solution Approach 1:
The control functions are segmented between two computers with clearly defined roles: the first computer handles communication and coordination, while the second computer dedicates resources to artificial intelligence and high-level functions. This segmentation frees up computation power for AI while managing system complexity through functional separation.
Solution Approach 2:
The first computer serves as an intermediary that manages the complexity of communication protocols and coordination tasks, allowing the second computer to focus on artificial intelligence without being burdened by communication overhead. This intermediary architecture balances computation power availability with manageable system complexity.
Data Source
AI summary
The present invention applies to a mobile robot which can have a human appearance and sophisticated functions enabling it to execute missions. To enable communications internal to the robot to be optimized and allow for a versatility that is both physical (possible substitution of parts of the robot) and software (replacement of programs to adapt it to new missions), an architecture with three computer levels is provided. The highest level comprises the intelligence of the robot which generates commands which are transmitted by an intermediate computer to low-level cards which control the sensors and actuators. Communication between the intermediate computer and the low-level cards is managed by at least one specific secure communication protocol.