Task adaptive dynamic computing power distribution method and system for heterogeneous robot

By constructing a task computing power library and a dynamic computing power allocation method, the problem of inflexible computing power management in heterogeneous robots is solved, thereby improving task execution efficiency and hardware adaptability, reducing energy consumption and improving endurance.

CN122363898APending Publication Date: 2026-07-10重庆中科汽车软件创新中心
2 Cites 0 Cited by

Patent Information

Application Number
CN202610463524.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-09
Publication Date
2026-07-10

AI Technical Summary

Technical Problem

Existing technologies for heterogeneous robots suffer from inflexible computing power management, failing to accurately match hardware status and task requirements, leading to resource waste, task failures, or system lag, and lacking an effective mechanism to cope with sudden high computing power demands.

Method used

By constructing a pre-set task computing power library, and combining task complexity, real-time performance, and hardware status, the system dynamically calculates and adapts computing power. When the matching is insufficient, it performs model pruning and knowledge distillation optimization, identifies non-core tasks for cloud-based distribution, and achieves adaptive dynamic computing power allocation for tasks.

Benefits of technology

It improves the task execution efficiency and hardware adaptability of heterogeneous robots, reduces energy consumption, enhances endurance and operational stability, and achieves efficient resource utilization.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention relates to the technical field of heterogeneous robots, in particular to a task self-adaption dynamic computing power distribution method and system for a heterogeneous robot. The method comprises the following steps: collecting typical task information of a heterogeneous robot, dividing tasks into different computing power demand levels according to task complexity, real-time requirements and data processing amount, and extracting key information to construct a preset task computing power library; receiving a robot task instruction, performing task analysis, setting a task real-time weight, collecting hardware state parameters of a robot in real time, and calculating adaptive calculation power; calculating the matching degree between the adaptive computing power and the computing power capable of being provided by the current hardware; when the matching degree is lower than a preset threshold value, structured channel pruning is carried out, and model fine adjustment is carried out through knowledge distillation; and when the computing power requirement cannot be met after pruning, identifying a non-core task and starting a cloud shunting process. According to the technical scheme, the task execution efficiency and the hardware adaptation capability of the heterogeneous robot can be improved.
Need to check novelty before this filing date? Find Prior Art