Year

2020

The idea

The mobile robotics system with pneumatic gripper hand and intelligent sensor technology stands for intuitive human-robot collaboration.

The objective

The BionicMobileAssistant with gripper hand is provided with artificial intelligence and moves autonomously in space. It can recognise objects, grasp them adaptively and work on them together with a human.

Technical data

  • Total weight: 22 kg
  • Battery: 48 V lithium polymer
  • Arm weight: 8 kg
  • Arm load capacity: 8 kg
  • Arm reach: 850 mm
  • Degrees of freedom (arm): 4
  • Degrees of freedom (hand): 11
  • Hand weight: 1.3 kg
  • Hand load capacity: max. 4 kg
  • Valve technology: 12 VTEM piezo cartridges

Structure

The robotic system comprises three subsystems: a mobile ballbot, an electric robotic arm and the BionicSoftHand 2.0, which was developed in 2019.

Mobility

The BionicMobileAssistant can manoeuvre in any direction. It balances on a sphere that is powered by three so-called Omniwheels. Since the ballbot only touches the ground at one point, it can navigate through narrow passages. Its planning and control algorithms are stored directly inside its corpus.

Dynamics

Under external influences, the robot sets the sphere in rotation and maintains its balance with maximum energy efficiency. By means of an inertial measurement unit and position encoders on its wheels, it registers its own movements and the relative inclination of the system. An optimisation program calculates how the robot and its arm must move to bring its gripper hand into the desired position, while also maintaining its stability.

Gripping

The BionicSoftHand 2.0 enables sensitive gripping. Tactile force sensors on the fingertips and on the palm and sides of the hand determine the hardness of the object grasped, enabling the pneumatic hand of the soft robotic system to adapt the force it exerts on it.

Object recognition

A depth camera for object detection is located in the wrist. The information is processed by a neural network that is trained in advance. In this way, the intelligent robotic hand can recognise and grasp objects even if they are partly obscured.

Material

BionicMobileAssistant