6.4.1. Force Control

Traditional control focuses on position or velocity. However, in tasks like grasping, polishing, or human-robot interaction, force control becomes essential.

In robotics, traditional control methods usually look at how to position or move the robot accurately by controlling its position (where it is) or its velocity (how fast it is moving). However, in certain tasks, like picking up an object (grasping), cleaning a surface (polishing), or interacting safely with humans, it's very important for the robot to control the amount of force it exerts on the object or the environment. This means that instead of just worrying about where the robot is or how fast it's moving, it's just as crucial to manage how hard it is pushing or pulling.

Imagine trying to pick up a delicate egg without breaking it. If you grip it too hard (too much force), it will crack. If you grip it too softly (too little force), it will slip from your fingers. Force control in robotics works similarly; it helps robots understand how to exert the right amount of pressure to safely grip an object.

Separates control into:

• Position control along unconstrained directions
• Force control along constrained directions

Requires knowledge of contact geometry.

Hybrid Position/Force Control is a method that combines both position and force control to handle different situations effectively. In areas where the robot can move freely without constraints (unconstrained directions), it will focus on position control, meaning it aims to place itself accurately at a specific point. Meanwhile, in situations where there are constraints, such as when pushing against a surface, the robot will focus on maintaining a constant force. To implement this, the robot must understand the shape and nature of the surfaces it interacts with, known as contact geometry, to apply the right amount of control in each type of direction.

Think of a person sweeping a floor. When they sweep the broom across a flat, open floor, they control the broom's position to get all the dirt. But if they hit a corner or an obstacle (like a piece of furniture), they need to adjust how much force they apply to the broom while still aiming to direct it correctly. The person uses both position (to guide the broom) and force (to push it effectively against the floor) just like a robot relies on hybrid control.

Models the robot as a virtual mass-spring-damper system:

F=Mx¨+Bx˙+Kx

• Impedance control: Robot behavior is specified by desired mechanical impedance.
• Admittance control: Especially for compliant robots, where force sensing is used to control position changes.

Widely used in cobots, rehabilitation robotics, and compliant manipulation.

Impedance and Admittance Control are advanced techniques that allow robots to interact more safely and effectively with their environments. By modeling the robot as a virtual mass-spring-damper, the system can respond to external forces dynamically, similar to how a real spring would respond when pushed or pulled. In Impedance Control, the robot is programmed to behave in a certain way based on the stiffness and damping properties defined by the designer. This is useful for tasks requiring delicate interaction. In Admittance Control, the focus is on how the robot will move or adjust its position in response to forces detected by sensors. This is particularly useful for more flexible robots used in cooperative tasks.

Imagine a trampoline. When a person jumps on it, the mat stretches and bends according to their weight and the force of their jump. Impedance control in a robot is like programming the trampoline to adjust how much it bounces back based on how hard someone jumps (the external force). Admittance control is like the trampoline's ability to lower its surface as more weight comes onto it, allowing for smoother landings. In rehabilitation settings, robots can use these controls to assist patients gently, responding to their need for support without causing any discomfort.

--