Cable conduit actuation provides a simple yet dexterous mode of power transmission for remote actuation. However, they are not preferred because of the nonlinearities arising from friction and cable compliance which lead to backlash type of behavior. Unlike most of the current research in backlash control which generally assumes no knowledge of one of the intermediate states, the controller design in this case can be significantly simplified if output feedback of the system is available. This paper uses a simple feedforward control law for backlash compensation. A novel smooth backlash inverse is proposed, which takes the physical limitations of the actuator in consideration, unlike other designs, and thus makes it more intuitive to use. Implementation of this inverse on physical systems can also improve the system performance over the theoretical exact inverse, as well as other existing smooth inverse designs. Improvement in the performance is shown through experiments on a robot arm of Laprotek surgical system as well as on an experimental setup using polymeric cables for actuation.

Many robotic systems, like surgical robots, robotic hands, and exoskeleton robots, use cable passing through conduits to actuate remote instruments. Cable actuation simplifies the design and allows the actuator to be located at a convenient location, away from the end effector. However, nonlinear frictions between the cable and the conduit account for major losses in tension transmission across the cable, and a model is needed to characterize their effects in order to analyze and compensate for them. Although some models have been proposed in the literature, they are lumped parameter based and restricted to the very special case of a single cable with constant conduit curvature and constant pretension across the cable only. This paper proposes a mathematically rigorous distributed parameter model for cable-conduit actuation with any curvature and initial tension profile across the cable. The model, which is described by a set of partial differential equations in the continuous time-domain, is also discretized for the effective numerical simulation of the cable motion and tension transmission across the cable. Unlike the existing lumped-parameter-based models, the resultant discretized model enables one to accurately simulate the partial-moving/partial-sticking cable motion of the cable-conduit actuation with any curvature and initial tension profile. The model is further extended to cable-conduit actuation in pull-pull configuration using a pair of cables. Various simulations results are presented to reveal the unique phenomena like backlash, cable slacking, interaction between the two cables, and other nonlinear behaviors associated with the cable conduits in pull-pull configuration. These results are verified by experiments using two dc motors coupled with a cable-conduit pair. The experimental setup has been prepared to emulate a typical cable-actuated robotic system. Experimental results are compared with the simulations and various implications are discussed.

Many surgical robots use cable-conduit pairs in a pull-pull configuration to actuate the instruments and transmit power into the patient’s body. Friction between the cable and the conduit makes the system nonlinear and accounts for major losses in tension transmission across the cable. This paper proposes an analytical model for a similar cable-conduit system and formulates the load transmission characteristics. The dynamic model uses discrete elements with friction losses and cable stretch calculated for each of the segments. The simulations predict backlash, cable slacking, and other nonlinear behavior. These results are verified with an experiment using two DC motors coupled with a cable-conduit pair. The drive motor is run in position control mode, while the load motor simulates a passive environment torsional spring. Experimental results are compared with the simulation and various implications are discussed.

Early detection and removal of small pulmonary nodules significantly improves long term survival rates for patients with lung cancer. To aid in the localization of these tumors during video assisted thoracoscopic surgery (VATS), a tactile imaging system (TIS) is presented. The system consists of a capacitive array sensor mounted on a minimally invasive surgical probe that is integrated with the thoracoscopic imaging. A vision-based algorithm localizes the probe in the live video and overlays a registered pseudo-color map of the measured pressure distribution on the streaming images. The surgeon can locate the hard nodules by scanning the tactile sensor head across the surface of the lung and observing the spatial variation in contact pressures caused by the elasticity differences in the underlying tissue. A validation experiment was conducted to compare the system to a current localization technique using a rigid rod. Results indicate that subjects could localize stiff lumps in lung phantoms more quickly and accurately using the TIS.

Natural orifice transluminal endoscopic surgery, or NOTES, allows for exceedingly minimally invasive surgery but has high requirements for the dexterity and force capabilities of the tools. An overview of the ViaCath System is presented. This system is a first generation teleoperated robot for endoluminal surgery and consists of a master console with haptic interfaces, slave drive mechanisms, and 6 degree-of-freedom, long-shafted flexible instruments that run alongside a standard gastroscope or colonoscope. The system was validated through animal studies. It was discovered that the devices were difficult to introduce into the GI tract and manipulation forces were insufficient. The design of a second generation system is outlined with improvements to the instrument articulation section and a steerable overtube. Results of basic evaluation tests performed on the tools are also presented.

This paper presents a system for relaying distributed tactile information from the finger tips of a remote manipulator to the finger tips of a human teleoperator. Tactile array sensors measure the pressure distribution at the contact between the robot hand and the grasped object. A computer samples this pressure signal and applies signal processing algorithms. The resulting signal drives a tactile shape display mounted on the master manipulator. This device raises an array of pins against the operator's finger tip skin to recreate the object shape. This system has been tested on a force-reflecting teleoperated hand. Experiments confirm the system's ability to convey significant contact information.