Robot Force-Field Approximation for Real-Time Complex Task Control

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In the potential method for robot control, calculating the artificial force field becomes computationally intensive and real-time processing is compromised when dealing with complex tasks.

Innovation Solution

An information processing device that acquires a task logical expression, samples a potential function from synthesized atomic potential functions, and calculates an approximate artificial force field to reduce computational load.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If atomic potential functions are synthesized according to the task logical expression to calculate the artificial force field, then the accuracy of the force field calculation is improved, but the calculation amount increases enormously and real-time processing cannot be ensured

Engineering Contradiction:
Improveaccuracy of artificial force field calculationVSAvoidreal-time processing capability
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent segments the calculation process into two distinct phases: an offline phase where multiple potential functions are pre-calculated and stored, and an online phase where only one potential function is sampled and used for real-time control. This segmentation allows the computationally intensive synthesis of atomic potential functions to be performed in advance, while real-time operation uses the pre-computed data, thus resolving the contradiction between calculation accuracy and real-time processing capability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary calculation and storage of multiple potential functions corresponding to different task logical expressions before actual robot control. By pre-computing the artificial force fields for various possible tasks and storing them, the system avoids performing these calculations during real-time control, ensuring that real-time processing can be executed quickly while maintaining high accuracy through the use of pre-computed potential functions.

Inventive Principle:
Principle #10Preliminary action

2Adaptability or versatility

If multiple potential functions are synthesized to handle complex robot tasks, then the task execution capability is improved, but the computational complexity increases

Engineering Contradiction:
Improvetask execution capabilityVSAvoidcomputational complexity
Core Design Contradiction:
Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent divides the computational workload by separating the complex synthesis of multiple potential functions from the simple sampling and application phases. The complex task of synthesizing atomic potential functions according to logical expressions is performed offline and stored, while online operation only requires sampling one pre-computed potential function, thus reducing online computational complexity while maintaining the ability to handle complex tasks.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent creates multiple copies of potential functions corresponding to different task logical expressions and stores them for later use. Instead of computing these functions repeatedly during real-time operation, the system prepares multiple copies in advance, allowing quick selection and application during actual robot control, thereby reducing computational complexity while preserving task execution capability.

Inventive Principle:
Principle #26Copying

Data Source

PatentUS12251842B2Information processing device, control method, and storage medium
Publication Date: 2025.03.18 NEC CORP
  • US12251842B2 patent drawing
  • US12251842B2 patent drawing
  • US12251842B2 patent drawing

AI summary

The information processing device 1A mainly includes an acquisition unit 51A, a sampling unit 54A, an artificial force field calculation unit 55A, and an output unit 56A. The acquisition unit 51A is configured to acquire a task logical expression in which an objective task to be performed by a robot is expressed by a combination of a plurality of atomic tasks. The sampling unit 54A is configured to perform sampling of one potential function from among a plurality of potential functions each of which is synthesized atomic potential functions, the atomic potential functions each corresponding to each of the atomic tasks. The artificial force field calculation unit 55A is configured to calculate, on a basis of the sampled potential function, an approximate artificial force field which approximates an artificial force field yielded when the atomic potential functions are synthesized according to the task logical expression. The output unit 56A is configured to output the approximate artificial force field.