Industrial robots are mostly used in industrial environments for handling and positioning tasks. Robots owe their precision to their rigid mechanical structure, which ensures that the joint angles set by the drives lead exactly to the end effector position calculated in the controller. However, making the mechanics as rigid as possible also has its disadvantages. It is particularly disadvantageous for tasks that require precise control of the handling forces that occur, for example when manipulating sensitive workpieces.

The rigid mechanics of industrial robots ensure a high bandwidth in force transmission, which means that high excitation frequencies are transmitted directly from the robot mechanics to the handled object. In addition, impulsive forces that act externally on the manipulated object are not sufficiently compensated. To put it vividly: picking up a raw egg from a fast-moving conveyor belt with a rigid robot without damaging it and placing it in a defined position on a rigid surface in a short space of time is an enormous challenge in terms of control technology.

A common solution strategy is therefore to introduce specific compliance into the handling process to reduce the bandwidth of force transmission and dampen high excitation frequencies. In the example, this would mean introducing damping materials such as rubber between the robot mechanics and the egg. However, these introduced compliant materials in turn result in inaccuracies, as forces acting on the handling object deform the compliant materials and thus change its position. In the above example, such forces correspond to the inertial forces that act on the egg during the robot movement and which increase the faster the robot accelerates.

A rigidly designed gripper mechanism therefore limits the maximum possible acceleration, as the acceleration forces that occur are not sufficiently damped and ultimately become too great for the sensitive surface of the ice. On the other hand, excessive compliance leads to strong position deviations even at low forces, so that no reliable statement can be made about the exact position of the manipulated object. A compromise must therefore always be found between positional accuracy and dynamics, which can be actively influenced by the compliance of the system.

The optimum solution in this example would be to adjust the compliance between the robot mechanics and the manipulation object during manipulation. This makes it possible to reduce acceleration and deceleration forces during start-up and braking of the manipulator by means of a high degree of compliance, whereas the swinging out of the manipulation object can be minimized by means of increased rigidity during set-down.

The Institute for Control Engineering of Machine Tools and Manufacturing Units (ISW) at the University of Stuttgart has developed a solution in the form of a robotic finger with adaptive flexibility based on shape memory alloys. So-called shape memory alloys (SMA) are used to be able to decouple the manipulated object from the rigid robot mechanics during handling if required. Due to their characteristic properties, these are suitable for the design of a gripper with variable stiffness. The design developed at the ISW is based on nickel-titanium alloys, so-called NiTiNol.

The mechanism of action of this shape memory alloy is based on a temperature- and stress-induced transformation of the material's crystal structure. In this process, monoclinic or orthorhombically distorted martensite transforms into densely packed body-centered cubic austenite. The change in volume of the material associated with the transformation leads to a contraction when heated, which can be used as a positioning movement for the robot finger. The reverse transformation as a result of cooling of the microstructure can be used accordingly for the counter-movement.

In simple terms, it can be assumed that the ratio between the martensite and austenite present determines the mechanical properties of the alloy. This ratio can be changed depending on the external loads by changing the temperature. The heat input required for this can be realized in a simple way using the law of electric heat by heating the alloy with an electric current. This allows the mechanical stiffness of the alloy to be controlled by the set electrical voltage.

Another advantage of shape memory alloys is their excellent force-to-weight ratio. In terms of gripper technology, this means that comparable gripping forces can be realized with significantly less weight at the end effector. For the developed concept, it is not necessary to attach heavy motors to the end effector of the robot, but the entire construction can be realized with lightweight materials. This helps to improve the dynamics of industrial robots, as inertial forces acting on the end effector, which affect all robot axes during movement and are amplified by the serial kinematics, are significantly reduced. The load capacity of the robot thus freed up can be used to increase the dynamics or to accommodate additional payloads.

The concept designed at the ISW is a prototype, whereby research into the properties of shape memory-based gripping mechanisms and the transfer of the concept into practice are at the forefront of the design. Possible areas of application are to be found particularly in the handling of sensitive workpieces. For example, when handling brittle materials such as ceramics, glass or workpieces with sensitive surfaces. M. Wnuk, T. Wenger, Dr. A. Lechler, Prof. Dr. A. Verl/as