Decentralized, sensor-based assistance system for ergonomic real-time monitoring and markerless object identification

Abstract

A portable assistance system for recording movement sequences and object parameters, characterized in that it consists of at least two spatially separated sensor units - a head-worn unit and an extremity-worn unit - each of which has inertial measurement units (IMUs) and communicates wirelessly with each other, wherein a computing unit calculates a difference angle (delta) between the direction of gaze and the direction of action of the extremity from the orientation data of both units in real time in order to quantify a torsional load on the user.

Description

Field of technology:

[0001] The invention relates to a wearable electronic system specifically designed for use in manual intralogistics and order picking. It serves for the preventive monitoring of ergonomic strain and the automated acquisition of process data (identification of packages) without the use of imaging camera systems. State of the art and disadvantages:

[0002] Current ergonomic monitoring solutions are mostly limited to individual sensors on the back or wrist that only measure local angles of inclination. However, they do not detect torsional movements (twisting of the spine) that arise from a discrepancy between the direction of gaze and the location of action.

[0003] Furthermore, the identification of packages in logistics currently mostly requires manual barcode scans or complex, camera-based AI systems. The latter are problematic from a data protection perspective (recognition of faces) and often fail in poor lighting conditions or extreme environmental conditions (e.g., deep-freeze storage, fog, dust). Purpose of the invention:

[0004] The invention is based on the objective of providing a system that: 1. Detects dangerous torsional movements by comparing multiple body axes in real time. 2. Identification of objects (e.g., boxes) is possible without using optical cameras. 3. Remains permanently functional under extreme climatic conditions (deep-freeze range down to -25 °C). Solution to the problem:

[0005] This task is solved by a decentralized sensor system consisting of at least one head unit and one arm unit that communicate wirelessly (e.g. via ESP-NOW protocol).

[0006] The invention utilizes a sensor fusion of LiDAR distance measurement (volume scan) and inertial measurement data (IMU: acceleration / gyroscope). By algorithmically comparing the orientation data of the head and arm, a difference angle (delta) is calculated, which serves as an indicator of ergonomic torsion.

[0007] At the same time, each moving object is assigned a dynamic signature (ID) generated from the captured volume (LiDAR grid) and the specific mass inertia (acceleration profile during lifting). Advantages of the invention: • Data protection compliance: Since no cameras are used, no biometric characteristics of employees are recorded. • Climate resilience: Thanks to special thermal encapsulation (EVA foam) and chemical sealing (conformal coating), use in deep-freeze storage is possible. • Process optimization: The creation of a digital 2D map (SLAM) is performed passively during the work process. 4. FIGURE DESCRIPTION

[0008] The invention is explained in more detail with reference to the accompanying drawings. These show: • Fig. 1: Schematic representation of the overall system on the user's body, showing the decentralized arrangement of head unit (10) and arm unit (20). • Fig. 2: Block diagram of the head unit, comprising the LiDAR sensor (11), the inertial measurement unit (12) and the radio module (13). • Fig. 3: Block diagram of the arm unit (master), comprising the computing unit (21), the data storage unit (22), the feedback display (23), the haptic sensor (24) and the inertial measurement unit of the extremity (25). • Fig. 4: Representation of the data flow, where LiDAR volume data (30) and acceleration data (31) are processed in a fusion calculation (32) to create a unique object ID (33). Reference symbol list: 10-head unit 11 LiDAR sensor 12 IMU heads 13 radio module 20 arm unit 21 computing unit 22 Data storage 23 Display 24 Haptic actuator 25 IMU arm 30 LiDAR data 31 Acceleration data 32 Data Fusion 33 Object ID

Claims

A portable assistance system for recording movement sequences and object parameters, characterized in that it consists of at least two spatially separated sensor units - a head-worn unit and an extremity-worn unit - each of which has inertial measurement units (IMUs) and communicates wirelessly with each other, wherein a computing unit calculates a difference angle (delta) between the direction of gaze and the direction of action of the extremity from the orientation data of both units in real time in order to quantify a torsional load on the user. System according to claim 1, characterized in that the head-mounted unit additionally has a LiDAR sensor (Light Detection and Ranging) which captures a multi-zone depth profile (e.g. 8x8 matrix) of the workspace, wherein these depth data are fused with the acceleration data of the extremity-mounted unit to assign a unique identification signature (ID) to moving objects based on their volume and inertia. System according to one of the preceding claims, characterized in that it has a ring buffer mechanism that continuously temporarily stores sensor data and only permanently saves the data of a defined time window (e.g. 10 seconds before the event) on a storage medium when a specific trigger (e.g. haptic sensor, pain threshold exceeded or detected anomalies) is triggered (e.g. event-driven forensic logging). System according to one of the preceding claims, characterized in that the electronic components are encapsulated by a combination of moisture-repellent conformal coating and a thermal insulation layer made of ethylene-vinyl acetate copolymer (EVA) in such a way that the inherent heat of the components and the body heat of the user are used to maintain the operating temperature of the energy storage device within the functional range even at ambient temperatures of up to -25 °C. System according to one of the preceding claims, characterized in that the decentralized units form a local wireless network independent of the infrastructure network and the data processing (in particular the calculation of the torsion and the object ID) takes place locally on the microcontrollers of the units (edge ​​computing), without requiring a permanent connection to a central server.

Images