Technology

Sensing the work environment

Nowadays, robots have taken over many one-sided repetitive work tasks that previously wore down workers. Relatively simple robotic tools work well when the tasks are well defined and the environment reamins constant.

For more challenging tasks, where multiple operations have to be performed in an unknown work environment, more advanced tools are needed. Graspian’s technology adds ”senses” to the end-of-arm tool enabling the robot to detect and adapt to various changing work environments.

Our core sensor technologies

Graspian’s end-of-arm tools include sensors for proximity, force and slippage.

Proximity: By measuring the distance to an object our tools are able to locate objects and feedback the location to the robot to guide the movements.

Touch force: By measuring the touch force applied when handling an object the tool can handle delicate objects like soft tomatoes or thin glass.

Object slippage: By measuring the object slippage the tool is able to detect if the object is about to glide out of the tool and react accordingly. Either by adding more force or putting the object down to get a better grip.

Advantages

Our sensors provide better and more consistant information about objects far beyond what simple vision cameras can offer. This enables us to locate and handle fragile, irregular and/or even slipping objects in unknown and changing work environments

Examples of fragile object are fruits and vegetables as well as glass and ceramics materials. Fruits and vegetables are also examples of irregular objects. With our sensors we are able to detect the exact location and shape of an object such that it can be picked up with the right amount of force to ensure a good grip without damaging the object through excessive force.

Slippage can happen when objects are coated by lubrication. By detecting slippage the tools will be able to rotate the object to get a better grip.

Step-by-step pick-and-place procedure

1. Object locationing

Apply computer vision to identify and locate an object, or use a priori knowledge about its location
Move the robot arm towards the object

2. Object picking

As the manipulator fingers approach the object, use tactile sensing to classify object compliance and ensure a good grip. A good grip is defined as one with sufficient force applied, even distribution of force from each robot finger, and avoiding to pick up an object close to edges and other irregular object features

3. Object lifting

When object is raised from it’s resting position, any unexpected motion caused by a poor grip or by slippage is being detected. If the motion happens instantly upon lift-off, then the procedure can be aborted by releasing the object.

4. Object moving

Other mission abortion techniques must be applied if the slippage or risk of droppage is detected when the object has been raised further from its resting position. Also, collision with the surrounding environment is detected by tactile sensors and abortion strategies can be put to practice.

5. Object placing

As the object approaches it’s destination, sensing of initial base contact is detected and a signal can be sent to the manipulator to release the object.
Alternatively, a human operator can tell a cobot (collaborative robot) to release an object, e.g. by pulling the object upwards. This will tell the robot system that the operator is taking over the picking control of the object.