Search

Compliant Movement Primitives (CMPs) showed good performance for a desirable behavior of robots to maintain low trajectory error while being compliant without knowing the dynamic model of the task. This framework uses the integral representation of reference trajectories in a feedback loop together with driving joint torques that represent the feed-forward control term. To achieve CMPs generalization, refer-ence trajectories (represented in the form of task space position trajectories) are encoded as Dynamic Movement Primitives (DMPs) while the feed-forward torques are learned through the Gaussian Process Regression (GPR) and are represented as a combination of radial basis functions. This paper extends the existing framework through the generalization of CMPs in bimanual settings that can concurrently achieve low trajectory errors in relative task space and compliant behavior in absolute task space. To achieve this behavior of the bimanual robotic system, the control terms derived from CMP framework are extended with the symmetric control approach. We show how the task-specific bimanual task dynamics can be learned and generalized to different task parameters that influence the task space trajectory and to a different load. Real-world results on a bimanual Kuka LWR-4 robots configuration confirms the usability of the extended framework.

Humans can reliably walk in dynamic and unstructured environments, simultaneously handle stairs and obstacles, as well as the transition from one gait to another. Despite the outstanding progress in the last decades, today's robots are still far from attaining that level of performance. This paper presents a new way of walk and stair climbing realization with a smooth transition between two types of gait. The gaits are composed of Reconfigurable Adaptive Motion Primitives (RAMPs), which serves as building blocks for any walking pattern. A human locomotion experiment is conducted to better understand how human approaches and positions the foot in front of the stairs with respect to the overall walk characteristics. These findings are used to compare the results of a simulation experiment with humanoid robot performing same locomotion as the human subject. It is shown that the robot's path or gait shape can be modified by the set of overall gait parameters that can be changed at any time instance. The robot can approach stairs with a variable number of half-steps, switch smoothly to stair climbing, and back to walking on flat surface, and modify walking speed and direction on-line as humans can do.

Humanoid robots are the most illustrative representatives of complex cyber-physical systems. They are high-dimensional and highly coupled non-linear systems. Moreover, walking humanoid robots are underactuated with friction-limited unilateral contacts with the ground. These facts justify the need for hard real-time control of a robot's motion. Even the smallest delay in calculating the reference motion or setting the output control variable can lead to a catastrophic outcome. For this reason, most of the existing robotic control systems are real-time and custom made, intended for a specific robot (or family of robots). There is no platform offering a general tool to design the model of versatile configurations of robots and to use it in the model-based control algorithms in hard real-time. In this paper, a software tool for modelling of an articulated system consisting of several open branched kinematic chains for a hard real-time OS is proposed. The model of the robot can possess an arbitrary number of links. The software is developed on top of the QNX Neutrino real-time operating system, thus providing the precise execution of time-critical events. In the experiment, the advantages of using the general model by modelling one real robotic configuration with multiple branched kinematic chains will be shown. The processing performance during the on-line motion planning will be tested on a real-time platform running on ARM Cortex-A8 processor. The results will be compared with similar software running in a non-real-time simulation environment.

The Government is deploying a growing number of radiation detectors at U.S. borders, ports of entry and other key locations, intended to assist in the detection of a potential terrorist's nuclear explosive device or radiological dispersal device. The most widely deployed radiation detector products today are radiation "pagers". Such products are unable to differentiate between natural radiation, and legitimate and clandestine radioactive materials, resulting in a disturbingly high incidence of false positives. As a result, there is a need for a portable device that can provide radionuclide identification and spatial localization. We report on the demonstration of a potentially low-cost, portable radiation sensor that provides radionuclide spectral information as well as spatial information and orientation about nuclear sources. Sensors based on this technology will allow personnel in the field to locate and positively identify potentially threatening radioactive materials while reducing the number of false alarms. Combined with isotopic identification software, it has the capability to differentiate radioactive sources as potentially dangerous or likely benign while also providing source direction, distance, and size information. It is based on high-resolution pixel array CZT detectors, new multichannel detector readout ASICs, coded mask apertures, and novel software. Data will be provided on device design and performance.

A low-cost, integrated ion implant-rapid thermal processor (RTP), SPC monitor has been developed to provide quick feedback for implant and RTP conditions critical to SPAWAR Systems Center, San Diego's (SSC San Diego) radiation hardened technologies. Low temperature solid phase epitaxy (SPE) was leveraged to achieve stable, reproducible responses at activation temperatures considerably below those used in conventional furnaces. Instabilities inherent to SPE processing required the design of experiment approach to determine implant and RTP conditions close to those used in SSC San Diego's baseline SOI technology, yet insensitive to room temperature drift (SPE-relaxation). A response methodology has been created to first isolate the cause of out-of-control signals to either the implanter or RTP system. Then, regression analyses on the response surface was used to quantify the drift in either dosimetry or temperature control.

Simultaneously achieving low trajectory errors and compliant control \emph{without} explicit models of the task was effectively addressed with Compliant Movement Primitives (CMP). For a single-robot task, this means that it is accurately following its trajectory, but also exhibits compliant behavior in case of perturbations. In this paper we extend this kind of behavior without explicit models to bimanual tasks. In the presence of an external perturbation on any of the robots, they will both move in synchrony in order to maintain their relative posture, and thus not exert force on the object they are carrying. Thus, they will act compliantly in their absolute task, but remain stiff in their relative task. To achieve compliant absolute behavior and stiff relative behavior, we combine joint-space CMPs with the well known symmetric control approach. To reduce the necessary feedback reaction of symmetric control, we further augment it with copying of a virtual force vector at the end-effector, calculated through the measured external joint torques. Real-world results on two Kuka LWR-4 robots in a bimanual setting confirm the applicability of the approach., 1st, unpolished version, 7 pages