Effective tactile sensing in prosthetic and robotic hands is crucial for improving the functionality of such hands and enhancing the user's experience. Thus, improving the range of tactile sensing capabilities is essential for developing versatile artificial hands. Multimodal tactile sensors called BioTacs, which include a hydrophone and a force electrode array, were used to understand how grip force, contact angle, object texture, and slip direction may be encoded in the sensor data. Findings show that slip induced under conditions of high contact angles and grip forces resulted in significant changes in both AC and DC pressure magnitude and rate of change in pressure. Slip induced under conditions of low contact angles and grip forces resulted in significant changes in the rate of change in electrode impedance. Slip in the distal direction of a precision grip caused significant changes in pressure magnitude and rate of change in pressure, while slip in the radial direction of the wrist caused significant changes in the rate of change in electrode impedance. A strong relationship was established between slip direction and the rate of change in ratios of electrode impedance for radial and ulnar slip relative to the wrist. Consequently, establishing multiple thresholds or establishing a multivariate model may be a useful method for detecting and characterizing slip. Detecting slip for low contact angles could be done by monitoring electrode data, while detecting slip for high contact angles could be done by monitoring pressure data. Predicting slip in the distal direction could be done by monitoring pressure data, while predicting slip in the radial and ulnar directions could be done by monitoring electrode data.

Although current urban search and rescue (USAR) robots are little more than remotely controlled cameras, the end goal is for them to work alongside humans as trusted teammates. Natural language communications and performance data are collected as a team of humans works to carry out a simulated search and rescue task in an uncertain virtual environment. Conditions are tested emulating a remotely controlled robot versus an intelligent one. Differences in performance, situation awareness, trust, workload, and communications are measured. The Intelligent robot condition resulted in higher levels of performance and operator situation awareness (SA).

Wearable robots including exoskeletons, powered prosthetics, and powered orthotics must add energy to the person at an appropriate time to enhance, augment, or supplement human performance. Adding energy while not being in sync with the user can dramatically hurt performance making it necessary to have correct timing with the user. Many human tasks such as walking, running, and hopping are repeating or cyclic tasks and a robot can add energy in sync with the repeating pattern for assistance. A method has been developed to add energy at the appropriate time to the repeating limit cycle based on a phase oscillator. The phase oscillator eliminates time from the forcing function which is based purely on the motion of the user. This approach has been simulated, implemented and tested in a robotic backpack which facilitates carrying heavy loads. The device oscillates the load of the backpack, based on the motion of the user, in order to add energy at the correct time and thus reduce the amount of energy required for walking with a heavy load. Models were developed in Working Model 2-D, a dynamics simulation software, in conjunction with MATLAB to verify theory and test control methods. The control system developed is robust and has successfully operated on a range of different users, each with their own different and distinct gait. The results of experimental testing validated the corresponding models.

Human running requires extensive training and conditioning for an individual to maintain high speeds (greater than 10mph) for an extended duration of time. Studies have shown that running at peak speeds generates a high metabolic cost due to the use of large muscle groups in the legs associated with the human gait cycle. Applying supplemental external and internal forces to the human body during the gait cycle has been shown to decrease the metabolic cost for walking, allowing individuals to carry additional weight and walk further distances. Significant research has been conducted to reduce the metabolic cost of walking, however, there are few if any documented studies that focus specifically on reducing the metabolic cost associated with high speed running. Three mechanical systems were designed to work in concert with the human user to decrease metabolic cost and increase the range and speeds at which a human can run.

The methods of design require a focus on mathematical modeling, simulations, and metabolic cost. Mathematical modeling and simulations are used to aid in the design process of robotic systems and metabolic testing is regarded as the final analysis process to determine the true effectiveness of robotic prototypes. Metabolic data, (VO2) is the volumetric consumption of oxygen, per minute, per unit mass (ml/min/kg). Metabolic testing consists of analyzing the oxygen consumption of a test subject while performing a task naturally and then comparing that data with analyzed oxygen consumption of the same task while using an assistive device.

Three devices were designed and tested to augment high speed running. The first device, AirLegs V1, is a mostly aluminum exoskeleton with two pneumatic linear actuators connecting from the lower back directly to the user's thighs, allowing the device to induce a torque on the leg by pushing and pulling on the user's thigh during running. The device also makes use of two smaller pneumatic linear actuators which drive cables connecting to small lever arms at the back of the heel, inducing a torque at the ankles. Device two, AirLegs V2, is also pneumatically powered but is considered to be a soft suit version of the first device. It uses cables to interface the forces created by actuators located vertically on the user's back. These cables then connect to the back of the user's knees resulting in greater flexibility and range of motion of the legs. Device three, a Jet Pack, produces an external force against the user's torso to propel a user forward and upward making it easier to run. Third party testing, pilot demonstrations and timed trials have demonstrated that all three of the devices effectively reduce the metabolic cost of running below that of natural running with no device.

Toward the ambitious long-term goal of a fleet of cooperating Flexible Autonomous Machines operating in an uncertain Environment (FAME), this thesis addresses several

critical modeling, design and control objectives for ground vehicles. One central objective was to show how off-the-shelf (low-cost) remote-control (RC) “toy” vehicles can be converted into intelligent multi-capability robotic-platforms for conducting FAME research. This is shown for two vehicle classes: (1) six differential-drive (DD) RC vehicles called Thunder Tumbler (DDTT) and (2) one rear-wheel drive (RWD) RC car called Ford F-150 (1:14 scale). Each DDTT-vehicle was augmented to provide a substantive suite of capabilities as summarized below (It should be noted, however, that only one DDTT-vehicle was augmented with an inertial measurement unit (IMU) and 2.4 GHz RC capability): (1) magnetic wheel-encoders/IMU for(dead-reckoning-based) inner-loop speed-control and outer-loop position-directional-control, (2) Arduino Uno microcontroller-board for encoder-based inner-loop speed-control and encoder-IMU-ultrasound-based outer-loop cruise-position-directional-separation-control, (3) Arduino motor-shield for inner-loop motor-speed-control, (4)Raspberry Pi II computer-board for demanding outer-loop vision-based cruise- position-directional-control, (5) Raspberry Pi 5MP camera for outer-loop cruise-position-directional-control (exploiting WiFi to send video back to laptop), (6) forward-pointing ultrasonic distance/rangefinder sensor for outer-loop separation-control, and (7) 2.4 GHz spread-spectrum RC capability to replace original 27/49 MHz RC. Each “enhanced”/ augmented DDTT-vehicle costs less than 􀀀175 but offers the capability of commercially available vehicles costing over 􀀀500. Both the Arduino and Raspberry are low-cost, well-supported (software wise) and easy-to-use. For the vehicle classes considered (i.e. DD, RWD), both kinematic and dynamical (planar xy) models are examined. Suitable nonlinear/linear-models are used to develop inner/outer-loopcontrol laws.

All demonstrations presented involve enhanced DDTT-vehicles; one the F-150; one a quadrotor. The following summarizes key hardware demonstrations: (1) cruise-control along line, (2) position-control along line (3) position-control along curve (4) planar (xy) Cartesian stabilization, (5) cruise-control along jagged line/curve, (6) vehicle-target spacing-control, (7) multi-robot spacing-control along line/curve, (8) tracking slowly-moving remote-controlled quadrotor, (9) avoiding obstacle while moving toward target, (10) RC F-150 followed by DDTT-vehicle. Hardware data/video is compared with, and corroborated by, model-based simulations. In short, many capabilities that are critical for reaching the longer-term FAME goal are demonstrated.

In-situ exploration of planetary bodies such as Mars or the Moon have provided geologists and planetary scientists a detailed understanding of how these bodies formed and evolved. In-situ exploration has aided in the quest for water and life-supporting chemicals. In-situ exploration of Mars carried out by large SUV-sized rovers that travel long distance, carry sophisticated onboard laboratories to perform soil analysis and sample collection. But their large size and mobility method prevents them from accessing or exploring extreme environments, particularly caves, canyons, cliffs and craters.

This work presents sub- 2 kg ball robots that can roll and hop in low gravity environments. These robots are low-cost enabling for one or more to be deployed in the field. These small robots can be deployed from a larger rover or lander and complement their capabilities by performing scouting and identifying potential targets of interest. Their small size and ball shape allow them to tumble freely, preventing them from getting stuck. Hopping enables the robot to overcome obstacles larger than the size of the robot.

The proposed ball-robot design consists of a spherical core with two hemispherical shells with grouser which act as wheels for small movements. These robots have two cameras for stereovision which can be used for localization. Inertial Measurement Unit (IMU) and wheel encoder are used for dead reckoning. Communication is performed using Zigbee radio. This enables communication between a robot and a lander/rover or for inter-robot communication. The robots have been designed to have a payload with a 300 gram capacity. These may include chemical analysis sensors, spectrometers and other small sensors.

The performance of the robot has been evaluated in a laboratory environment using Low-gravity Offset and Motion Assistance Simulation System (LOMASS). An evaluation was done to understand the effect of grouser height and grouser separation angle on the performance of the robot in different terrains. The experiments show with higher grouser height and optimal separation angle the power requirement increases but an increase in average robot speed and traction is also observed. The robot was observed to perform hops of approximately 20 cm in simulated lunar condition. Based on theoretical calculations, the robot would be able to perform 208 hops with single charge and will operate for 35 minutes. The study will be extended to operate multiple robots in a network to perform exploration. Their small size and cost makes it possible to deploy dozens in a region of interest. Multiple ball robots can cooperatively perform unique in-situ science measurements and analyze a larger surface area than a single robot alone on a planet surface.

To ensure system integrity, robots need to proactively avoid any unwanted physical perturbation that may cause damage to the underlying hardware. In this thesis work, we investigate a machine learning approach that allows robots to anticipate impending physical perturbations from perceptual cues. In contrast to other approaches that require knowledge about sources of perturbation to be encoded before deployment, our method is based on experiential learning. Robots learn to associate visual cues with subsequent physical perturbations and contacts. In turn, these extracted visual cues are then used to predict potential future perturbations acting on the robot. To this end, we introduce a novel deep network architecture which combines multiple sub- networks for dealing with robot dynamics and perceptual input from the environment. We present a self-supervised approach for training the system that does not require any labeling of training data. Extensive experiments in a human-robot interaction task show that a robot can learn to predict physical contact by a human interaction partner without any prior information or labeling. Furthermore, the network is able to successfully predict physical contact from either depth stream input or traditional video input or using both modalities as input.

Sports activities have been a cornerstone in the evolution of humankind through the ages from the ancient Roman empire to the Olympics in the 21st century. These activities have been used as a benchmark to evaluate the how humans have progressed through the sands of time. In the 21st century, machines along with the help of powerful computing and relatively new computing paradigms have made a good case for taking up the mantle. Even though machines have been able to perform complex tasks and maneuvers, they have struggled to match the dexterity, coordination, manipulability and acuteness displayed by humans. Bi-manual tasks are more complex and bring in additional variables like coordination into the task making it harder to evaluate.

A task capable of demonstrating the above skillset would be a good measure of the progress in the field of robotic technology. Therefore a dual armed robot has been built and taught to handle the ball and make the basket successfully thus demonstrating the capability of using both arms. A combination of machine learning techniques, Reinforcement learning, and Imitation learning has been used along with advanced optimization algorithms to accomplish the task.

To achieve the ambitious long-term goal of a feet of cooperating Flexible Autonomous

Machines operating in an uncertain Environment (FAME), this thesis addresses several

critical modeling, design, control objectives for rear-wheel drive ground vehicles.

Toward this ambitious goal, several critical objectives are addressed. One central objective of the thesis was to show how to build low-cost multi-capability robot platform

that can be used for conducting FAME research.

A TFC-KIT car chassis was augmented to provide a suite of substantive capabilities.

The augmented vehicle (FreeSLAM Robot) costs less than $500 but offers the capability

of commercially available vehicles costing over $2000.

All demonstrations presented involve rear-wheel drive FreeSLAM robot. The following

summarizes the key hardware demonstrations presented and analyzed:

(1)Cruise (v, ) control along a line,

(2) Cruise (v, ) control along a curve,

(3) Planar (x, y) Cartesian Stabilization for rear wheel drive vehicle,

(4) Finish the track with camera pan tilt structure in minimum time,

(5) Finish the track without camera pan tilt structure in minimum time,

(6) Vision based tracking performance with different cruise speed vx,

(7) Vision based tracking performance with different camera fixed look-ahead distance L,

(8) Vision based tracking performance with different delay Td from vision subsystem,

(9) Manually remote controlled robot to perform indoor SLAM,

(10) Autonomously line guided robot to perform indoor SLAM.

For most cases, hardware data is compared with, and corroborated by, model based

simulation data. In short, the thesis uses low-cost self-designed rear-wheel

drive robot to demonstrate many capabilities that are critical in order to reach the

longer-term FAME goal.

The Basilisk lizard is known for its agile locomotion capabilities on granular and aquatic media making it an impressive model organism for studying multi-terrain locomotion mechanics. The work presented here is aimed at understanding locomotion characteristics of Basilisk lizards through a systematic series of robotic and animal experiments. In this work, a Basilisk lizard inspired legged robot with bipedal and quadrupedal locomotion capabilities is presented. A series of robot experiments are conducted on dry and wet (saturated) granular media to determine the effects of gait parameters and substrate saturation, on robot velocity and energetics. Gait parameters studied here are stride frequency and stride length. Results of robot experiments are compared with previously obtained animal data. It is observed that for a fixed robot stride frequency, velocity and stride length increase with increasing saturation, confirming the locomotion characteristics of the Basilisk lizard. It is further observed that with increasing saturation level, robot cost of transport decreases. An identical series of robot experiments are performed with quadrupedal gait to determine effects of gait parameters on robot performance. Generally, energetics of bipedal running is observed to be higher than quadrupedal operation. Experimental results also reveal how gait parameters can be varied to achieve different desired velocities depending on the substrate saturation level. In addition to robot experiments on granular media, a series of animal experiments are conducted to determine and characterize strategies

exhibited by Basilisk lizards when transitioning from granular to aquatic media.