GM Gripper Challenge
GM, in partnership with MassRobotics, is holding a Gripper Challenge.
We are seeking startups, academia and integrators to develop a proof-of-concept universal end effector capable of grasping and holding a variety of sheet metal components for purposes of precise robotic assembly.
The end effector must be capable of holding parts firmly enough for precision alignment with parts to enable precision assembly tasks.
Key performance metrics for concept selection include: geometric flexibility, holding force, cycle time, reliability and simplicity.
The end effector must:
- Hold 12+ different parts
- Hold up to 25lbs
- Be capable of precision alignment with parts
- Use sensor-based robotic control
- Endure spot welding and other processing while gripping
- Require minimal integration and be self-contained
Performance will be judged on:
- Holding force
- Speed (Cycle time < 5 sec)
GM will provide sample parts. Entrants will retain IP.
Frequently Asked Questions
Applications are primarily ferrous. However a gripper that can handle non-ferrous (e.g. aluminum or composite) parts will be viewed more preferably because it can handle a wider variety of parts, but is not a requirement.
Weight of objects?
Our primary concern is with smaller sheet metal parts (e.g. up to 5 lbs.). However, a gripper that can handle larger and heavier parts will be viewed more preferably because it can handle a wider variety of parts. Beyond weight of the object, the gripper must be capable of holding the part in fixed position when external loads are applied to the part as part of the manufacturing process. Maximum external loads would be up to 25 lbs.
Relative size of end effector?
The gripper shouldn’t be significantly larger than the largest part held by the gripper. The combined weight of the gripper and part should not exceed ~180 kg and the gripper to part size ratio should be between 1.5-2:1. That is for example to grasp parts of 0.5 cu ft a gripper should be below 1.5 cu ft in volume. The size of the gripper should be minimized to enable the maximum movement without collision with other robots or equipment in a manufacturing cell. The gripper should grip the part only from one side (leaving the opposite side completely open for processing, or placing against a surface, or to view the surface of the part with a vision system. The gripper should allow for maximum accessibility to the part to allow for processing (especially resistance spot welding or applying adhesives and sealants). Grippers that completely envelop the part are less desirable because they restrict the ability to access the part for processing, machine vision line of sight, or part placement against surfaces.
Are there defined zero points? How many?
Most parts include defined datums with higher precision. These are usually defined by 2 round holes and a surface, or a hole and slot and a surface. A gripper may use these features for precision location. A gripper might use any number of positive (e.g.: surfaces) or negative (e.g.: holes) features to constrain the position and orientation of the part relative to the robot. However, a gripper that uses a minimal number of mechanical devices (e.g.: 3 pins) to hold the part firmly will be preferred.
Our goal would be able to achieve position tolerances on the order of 0.1mm.
Are parts picked up by the robot or staged?
We have different kinds of applications. It would be best if the gripper can be used for bin picking, picking from a presentation nest, or hand-off from another robot. However, for the application used to validate the design of the gripper the parts will be presented on a part nest at a fixed position and orientation. The parts will be manually loaded to the presentation next.
The gripper should be IP54 or above.
Can intermediary parts be used?
The use of intermediate parts that would make the design dependent to specific part geometries should be avoided. The gripper should be designed to enable the reconfiguration to grasp different parts using actuation (e.g.: servo-actuated features) instead of intermediate parts.
Electric and/or pneumatic are allowable power sources. 24 volts is our preference for electric. 80psi air pressure is available. For this PoC, it’s fine if the gripper is powered using 110V. It doesn’t need to be powered from the robot directly. However, in the long run, our goal would be to power the gripper through cables and lines that run through standard robot dress packages.
Price point desired?
In general, a flexible gripper would be more valuable to us than a dedicated gripper. However, a flexible gripper might be purchased in higher quantities since the same design can be used in many applications. Higher purchase quantities might lead to lower cost due to economies of scale.
Detection + software integration?
Inclusion of sensing to detect part presence and/or other fault conditions (e.g.: part dropped, part loose) would be viewed favorably, but not necessary in a PoC. The controls software can run on a separate PC or device for the PoC gripper.