Abstract² From exploring planets to cleaning homes, the reach and versatility of robotics is vast. The integration of actuation, sensing and control makes robotics systems powerful, but complicates their simulation. This paper introduces a versatile, scalable, yet powerful general-purpose robot simulation framework called V-REP.The paper discusses the utility of a portable and flexible simulation framework that allows for direct incorporation of various control techniques. This renders simulations and simulation models more accessible to a general-public, by reducing the simulation model deployment complexity. It also increases productivity by offering built-in and ready-to-use functionalities, as well as a multitude of programming approaches.This allows for a multitude of applications including rapid algorithm development, system verification, rapid prototyping, and deployment for cases such as safety/remote monitoring, training and education, hardware control, and factory automation simulation.

Abstract. From exploring planets to cleaning homes, the reach and versatility of robotics is vast. The integration of actuation, sensing and control makes robotics systems powerful, but complicates their simulation. This paper introduces a modular and decentralized architecture for robotics simulation. In contrast to centralized approaches, this balances functionality, provides more diversity, and simplifies connectivity between (independent) calculation modules. As the Virtual Robot Experimentation Platform (V-REP) demonstrates, this gives a smallfootprint 3D robot simulator that concurrently simulates control, actuation, sensing and monitoring. Its distributed and modular approach are ideal for complex scenarios in which a diversity of sensors and actuators operate asynchronously with various rates and characteristics. This allows for versatile prototyping applications including systems verification, safety/remote monitoring, rapid algorithm development, and factory automation simulation.

Using an instrumented surgical tool, high-precision recordings of hand tremor were taken during vitreoretinal microsurgery. The data obtained using a compact, custom sixdegree-of-freedom inertial sensing module were filtered and analyzed to characterize the physiological hand tremor of the surgeon. Tremor during the most delicate part of the procedure was measured at a vector magnitude of 38 pm rms. Nontremulous, lower-frequency components of instrument movement were also characterized. The data collected provide an important baseline for design specification and performance evaluation of engineered microsurgical devices.

Abstract-Ohio State and Stanford Universities are cooperating to construct and artificial quadruped. This quadruped has a rest leg length of 0.68m and weighs 70kg. The body is made of four single-leg modules. Each module houses an articulated 3-dof leg. The articulated leg structure uses a set of mechanical springs to store energy. The stiffness of the leg, from hip to foot, is highly nonlinear. The leg initially stiffens as the leg is compressed the first 1/3 of its range, then remains roughly constant for the remainder. This stiffening allows the leg to maintain a relatively constant working length. A cable is used to flex the knee of the quadruped. As the cable is pulled, the knee bends, tensioning the energy storage springs. This cable makes about a ¼ turn over a pulley which is concentric with the hip axis, ½ turn around a pulley at the ankle and then ¾ turn back around a second hip pulley in a direction opposite the first hip pulley. This arrangement decouples cable tension and hip torque. The non-anchored end of the cable is tensioned by a capstan. This capstan uses unique geometry to decrease holding torque and to release the cable instantly. The current top speed of this leg in a 2-dimensional test is 4.15m/s. This leg has been controlled using two control systems, both implemented on an embedded controller attached to the leg. An articulated leg presents control challenges not seen when trying to control a prismatic leg. The first controller tested is a heuristic algorithm whose parameters are updated in real time by the LevenbergMarquardt learning method. This controller is similar to the controllers used by Raibert[24], with modifications to allow for leg asymmetry. The second controller tested is a direct adaptive fuzzy controller. The fuzzy controller consists of a rule base, inference mechanism, fuzzification interface, and defuzzification interface. Fuzzification starts by mapping an input (body velocity, desired change in velocity, height) into one or more membership functions.The inference mechanism then determines the applicability of each rule to the current inputs. Defuzzification combines the recommendations of each rule in an output based upon rule certainties. The adaptation mechanism modifies the rule output centers to correct velocity errors. The adaptation mechanism gains are experimentally tuned. Once tuned, the controller quickly adapts to leg changes. Both control systems were tested with and without adaptation. Both systems more accurately tracked desired velocity with adaptation. Accurate, high resolution, high speed body attitude sensing is essential for successful quadruped operation.No existing solutions meet the project requirements. An adequate sensor system is being developed in cooperation with a commercial vendor. Initial results show that accurate position tracking is possible with currently available MEMS inertial sensors.

For agile, accurate autonomous robotics, it is desirable to plan motion in the presence of uncertainty. The Partially Observable Markov Decision Process (POMDP) provides a principled framework for this. Despite the tremendous advances of POMDP-based planning, most can only solve problems with a small and discrete set of actions. This paper presents General Pattern Search in Adaptive Belief Tree (GPS-ABT), an approximate and online POMDP solver for problems with continuous action spaces. Generalized Pattern Search (GPS) is used as a search strategy for action selection. Under certain conditions, GPS-ABT converges to the optimal solution in probability. Results on a box pushing and an extended Tag benchmark problem are promising.