Feedback signals from a transducer in an artificial hand or robot gripper are essential for the successful manipulation of objects. A series of different control strategies have been developed to improve the performance of an anthropomorphic prosthesis by detecting when the object slips. The methods are designed to be independent and so their combination improves the system's total performance by removing the false detection of slides whilst increasing sensitivity. The principal method of transduction uses a novel combined touch and slip sensor to detect the contact force as well as the vibrations set up when the object slips within the hand's grasp. Static and dynamic characteristics of the sensor are described. The frequency response of the sensor and the electronic post-processing vary for differing conditions. The underlying factors of the signal generation and detection are sufficiently common to make recommendations concerning the requirements for the signal processing of the output possible. While the target application for this method is prosthetics, similar techniques can be applied to more general classes of robotic manipulation. These methods would improve a robot's response because the device does not require detailed recognition of a target object or its orientation prior to handling.