Our vision is to make intelligent robots available to everyone – to help us with manual work in manufacturing, in the office and at home. This goal necessitates robots with greatly increased sensitivity and perception capabilities, but simplified and cheaper hardware.
RoVi offers an innovative and cost-effective software-based sensor solution to achieve this goal. Our product features rich sensory feedback, providing measurements of joint positions of the robot arm, as well as force, torque and tactile data at the end-effector.
Sensing is performed remotely, using a standard camera connected to our vision software. Complex and costly hardware sensors are replaced by our software, reducing hardware complexity and weight, as well as costs of components, assembly and maintenance.
Joint position sensors or encoders are indispensible for any robot arm. Our technology determines the joint states remotely, using only a camera and an image processing software. The software accurately localizes patterns on the robot arm, indicated in green in the picture, and calucates calculates the joint positions from these visual measurments.
These sensors measure 3D force and 3D torque, usually between the robot arm and the tool. They allow for force control and teaching of motions. Our solution relies on a very simple passive, elastic structure, which bends when forces are applied. This bending is observed by the camera and converted to a force/torque signal by our image processing software.
Tactile sensors provide a pressure or contact profile, similar to touch we feel on our skin. They offer rich feedback about the grasping process and are thus essential for adaptive and sensitive grasping processes. RoVi's camera-based solution requires only a passive and low-cost foam bar on the fingers, which deforms on pressure. Our image processing software derives a tactile signal from the observed deformation.
The following shows some exemplary use cases and applications for robots using our sensing system. Do not hesitate to tell us about your application! We will provide a quick evaluation of what our technology can do to solve your automation problem!
Our camera-based sensing technology enables lightweight, low-cost and mobile robot arms. They fit on any desk or workbench to assist a human worker. With integrated computer vision, they are easily adaptable and can detect objects.
Take your grasping applications to the next level! Handle objects of different shapes, rigidity and material with a single versatile and self-adapting gripper. With our tactile sensor, you get rich feedback about the grasped object and the grasping state.
Commission small parts such as screws, fittings or plastic part in exact quanities for packagaing or for subsequent worksteps. The robot can do this step even right on the workbench using our versatile grasping system, or any another manipulator.
Monitor the status of your automation systems by adding our software-based sensors. Data and statistics are processed on-site or in the cloud and presented on any web browser.
Package parts or components provided by a machine or a human worker. Our camera-based system flexibly localizes new parts, picks them and places them to a free spot.
This demo video shows the fully camera-based control of a "sensorless" robot arm.
Shows how we build a sensitive gripper from low-cost, 3D-printed parts and presents the rich feedback provided by our tactile sensor.
The robot in this video opens a beer bottle using a 6D camera-based force/torque sensor.
RoVi was founded by three engineers as a spin-off from Technical University of Munich (TUM). Nicolas and Clemens have been working together for several years in the field of robotics. Our camera-based sensing technology mainly comes from Nicolas' research in computer vision and visuo-haptic perception. Clemens' research focus was in haptics and telepresence. Stefan is an experienced electronics and hardware engineer.