Four-legged robot multi-modal perception control device suitable for inspection

By integrating a foot vibration sensing module and an energy management module into a multimodal sensing and control device, the gait and sensing parameters of the quadruped robot are dynamically adjusted, resolving the contradiction between vibration sampling quality and energy consumption. This enables the inspection system to achieve efficient detection and long endurance, and improves the inspection capability in complex terrain.

CN122331597APending Publication Date: 2026-07-03JIEYAO INTELLIGENT TECHNOLOGY (SUZHOU) CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
JIEYAO INTELLIGENT TECHNOLOGY (SUZHOU) CO LTD
Filing Date
2026-04-20
Publication Date
2026-07-03

AI Technical Summary

Technical Problem

In existing technologies, quadruped robots face a contradiction between vibration sampling quality and energy consumption in inspection tasks. High-quality vibration sampling requires low speed and long ground contact time, which consumes more energy. Low-energy fast walking gait leads to a decrease in vibration signal quality, affecting the accuracy of anomaly detection. It is impossible to ensure both detection accuracy and endurance under limited energy constraints.

Method used

Employing a multimodal sensing and control device, integrating a foot vibration sensing module, an energy management module, and a vibration-energy co-controller, the device dynamically adjusts gait parameters and sensing system parameters through a unified optimization objective function to achieve an adaptive balance between vibration quality and energy consumption. This includes a high-bandwidth vibration sensor array, a six-dimensional force sensor, a PVDF piezoelectric film, a fuselage inertial measurement unit, and a lightweight convolutional autoencoder network. Combined with a lightweight neural network and a multi-objective optimizer, it monitors and adjusts walking speed, stride length, ground contact time, and sensor sampling frequency in real time.

Benefits of technology

During the inspection process, a dynamic balance between vibration signal quality and energy consumption was achieved, which improved the overall reliability and detection accuracy of the inspection system, extended its battery life, and enabled it to continuously learn and improve its performance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122331597A_ABST
    Figure CN122331597A_ABST
Patent Text Reader

Abstract

This invention discloses a multimodal perception and control device for quadruped robots suitable for inspection, belonging to the field of robot control technology. The device includes a foot vibration sensing module, an energy management module, a vibration-energy collaborative controller, a gait control module, and a perception system scheduling module. The foot vibration sensing module collects foot-ground contact vibration spectrum data during the ground contact phase; the energy management module monitors the remaining battery energy and the power consumption of each component in real time; the vibration-energy collaborative controller, with the inspection task completion rate as a unified objective, calculates vibration quality indicators based on vibration data, calculates the remaining energy budget based on remaining energy and power consumption, and outputs gait and perception parameter adjustment commands; the gait control module adjusts walking speed and motor torque curves; and the perception system scheduling module adjusts the sensor sampling frequency and resolution. This invention achieves an adaptive balance between vibration sampling quality and energy consumption, extending the effective inspection time while ensuring anomaly detection accuracy.
Need to check novelty before this filing date? Find Prior Art

Citation Information

Patent Citations

  • Four-footed inspection robot space positioning method and system based on gravity center sensing

    CN120869163A

  • Four-legged robot control method and system based on transformer reinforcement learning architecture

    CN121115515B