Controlling the estimation bias in deep reinforcement learning problems with sparse rewards: Towards robust robotic object manipulation learning

Many recent robot learning problems, real and simulated, were addressed using deep reinforcement learning. The developed policies can deal with high-dimensional, continuous state and action spaces, and can also incorporate machine-generated or human demonstration data. A great number of them depend on state-action value estimates, especially the...

Controlling grip force by maintaining a constant frictional safety margin to improve robotic grasping

Manipulating soft and fragile objects is a challenging task in robotic grasping. The key challenge for robotic grasping is to exert enough grip force to prevent slipping while being gentle enough to prevent damage to an object. Existing grippers used for processes like automatic harvesting of fruits, either apply excessive grip force leading to...

Statically balanced singular-friction locking

Locking mechanisms are used in a wide range of applications including energy saving mechanisms in the field of robotics. In this thesis a new class of locking mechanisms is investigated: a combination between a singular- and a friction locking mechanism (singular-friction locker) which is statically balanced with one additional spring. This...