The letter presents a force-tracking impedance controller granting a free-overshoots contact force (mandatory performance for many critical interaction tasks such as polishing) for partially unknown interacting environments (such as leather or hard-fragile materials). As in many applications, the robot has to gently approach the target environment (whose position is usually not well-known), then execute the interaction task. Therefore, the algorithm has been designed to deal with both the free space approaching motion (phase a.) and the succeeding contact task (phase b.) without switching from different control logics. Control gains have to be properly calculated for each phase in order to achieve the target force tracking performance (i.e., free-overshoots contact force). In detail, phase a. control gains are optimized based on the impact collision model to minimize the force error during the following contact task, while phase b. control gains are analytically calculated based on the solution of the LQR optimal control problem. The analytical solution grants the continuous adaptation of the control gains during the contact phase on the estimated value of the environment stiffness (obtained through an on-line extended Kalman filter). A probing task has been carried out to validate the performance of the control with partially unknown contact environment properties. Results show the avoidance of force overshoots and instabilities.

The paper focuses on industrial interaction robotics tasks, investigating a control approach involving multiples learning levels for training the manipulator to execute a repetitive (partially) changeable task, accurately controlling the interaction. Based on compliance control, the proposed approach consists of two main control levels: 1) iterative friction learning compensation controller with reinforcement and 2) iterative force-tracking learning controller with reinforcement. The learning algorithms rely on the iterative learning and reinforcement learning procedures to automatize the controllers parameters tuning. The proposed procedure has been applied to an automotive industrial assembly task. A standard industrial UR 10 Universal Robot has been used, equipped by a compliant pneumatic gripper and a force/torque sensor at the robot end-effector.

This paper presents a learning-based method that uses simulation data to learn an object manipulation task using two model-free reinforcement learning (RL) algorithms. The learning performance is compared across on-policy and off-policy algorithms: Proximal Policy Optimization (PPO) and Soft Actor-Critic (SAC). In order to accelerate the learning process, the fine-tuning procedure is proposed that demonstrates the continuous adaptation of on-policy RL to new environments, allowing the learned policy to adapt and execute the (partially) modified task. A dense reward function is designed for the task to enable an efficient learning of the agent. A grasping task involving a Franka Emika Panda manipulator is considered as the reference task to be learned. The learned control policy is demonstrated to be generalizable across multiple object geometries and initial robot/parts configurations. The approach is finally tested on a real Franka Emika Panda robot, showing the possibility to transfer the learned behavior from simulation. Experimental results show 100% of successful grasping tasks, making the proposed approach applicable to real applications.