Human-robot planar co-manipulation of extended objects: data-driven models and control from human-human dyads.

Human teams are able to easily perform collaborative manipulation tasks. However, simultaneously manipulating a large extended object for a robot and human is a difficult task due to the inherent ambiguity in the desired motion. Our approach in this paper is to leverage data from human-human dyad experiments to determine motion intent for a physical human-robot co-manipulation task. We do this by showing that the human-human dyad data exhibits distinct torque triggers for a lateral movement. As an alternative intent estimation method, we also develop a deep neural network based on motion data from human-human trials to predict future trajectories based on past object motion. We then show how force and motion data can be used to determine robot control in a human-robot dyad. Finally, we compare human-human dyad performance to the performance of two controllers that we developed for human-robot co-manipulation. We evaluate these controllers in three-degree-of-freedom planar motion where determining if the task involves rotation or translation is ambiguous.

Design Optimization for Rough Terrain Traversal Using a Compliant, Continuum-Joint, Quadruped Robot.

Legged robots have the potential to cover terrain not accessible to wheel-based robots and vehicles. This makes them better suited to perform tasks such as search and rescue in real-world unstructured environments. In addition, pneumatically-actuated, compliant robots may be more suited than their rigid counterparts to real-world unstructured environments with humans where unintentional contact or impact may occur. In this work, we define design metrics for legged robots that evaluate their ability to traverse unstructured terrain, carry payloads, find stable footholds, and move in desired directions. These metrics are demonstrated and validated in a multi-objective design optimization of 10 variables for a 16 degree of freedom, pneumatically actuated, continuum joint quadruped. We also present and validate approximations to preserve numerical tractability for any similar high degree of freedom optimization problem. Finally, we show that the design trends uncovered by our optimization hold in two hardware experiments using robot legs with continuum joints that are built based on the optimization results.

Performance Metrics for Fluidic Soft Robot Rotational Actuators.

The field of soft robotics is continuing to grow as more researchers see the potential for robots that can safely interact in unmodeled, unstructured, and uncertain environments. However, in order for the design, integration, and control of soft robotic actuators to develop into a full engineering methodology, a set of metrics and standards need to be established. This paper attempts to lay the groundwork for that process by proposing six soft robot actuator metrics that can be used to evaluate and compare characteristics and performance of soft robot actuators. Data from eight different soft robot rotational actuators (five distinct designs) were used to evaluate these soft robot actuator metrics and show their utility. Additionally we provide a simple case study as an example of how these metrics can be used to evaluate soft robot actuators for a designated task. While this paper does not claim to present a comprehensive list of all possible soft robot actuator metrics, the metrics presented can 1) be used to initiate the development and comparison of soft robot actuators in an engineering framework and 2) start a broader discussion of which metrics should be standardized in future soft robot actuator research.

Improved Continuum Joint Configuration Estimation Using a Linear Combination of Length Measurements and Optimization of Sensor Placement.

This paper presents methods for placing length sensors on a soft continuum robot joint as well as a novel configuration estimation method that drastically minimizes configuration estimation error. The methods utilized for placing sensors along the length of the joint include a single joint length sensor, sensors lined end-to-end, sensors that overlap according to a heuristic, and sensors that are placed by an optimization that we describe in this paper. The methods of configuration estimation include directly relating sensor length to a segment of the joint's angle, using an equal weighting of overlapping sensors that cover a joint segment, and using a weighted linear combination of all sensors on the continuum joint. The weights for the linear combination method are determined using robust linear regression. Using a kinematic simulation we show that placing three or more overlapping sensors and estimating the configuration with a linear combination of sensors resulted in a median error of 0.026% of the max range of motion or less. This is over a 500 times improvement as compared to using a single sensor to estimate the joint configuration. This error was computed across 80 simulated robots of different lengths and ranges of motion. We also found that the fully optimized sensor placement performed only marginally better than the placement of sensors according to the heuristic. This suggests that the use of a linear combination of sensors, with weights found using linear regression is more important than the placement of the overlapping sensors. Further, using the heuristic significantly simplifies the application of these techniques when designing for hardware.

Trends in Haptic Communication of Human-Human Dyads: Toward Natural Human-Robot Co-manipulation.

In this paper, we analyze and report on observable trends in human-human dyads performing collaborative manipulation (co-manipulation) tasks with an extended object (object with significant length). We present a detailed analysis relating trends in interaction forces and torques with other metrics and propose that these trends could provide a way of improving communication and efficiency for human-robot dyads. We find that the motion of the co-manipulated object has a measurable oscillatory component. We confirm that haptic feedback alone represents a sufficient communication channel for co-manipulation tasks, however we find that the loss of visual and auditory channels has a significant effect on interaction torque and velocity. The main objective of this paper is to lay the essential groundwork in defining principles of co-manipulation between human dyads. We propose that these principles could enable effective and intuitive human-robot collaborative manipulation in future co-manipulation research.

Using First Principles for Deep Learning and Model-Based Control of Soft Robots.

Model-based optimal control of soft robots may enable compliant, underdamped platforms to operate in a repeatable fashion and effectively accomplish tasks that are otherwise impossible for soft robots. Unfortunately, developing accurate analytical dynamic models for soft robots is time-consuming, difficult, and error-prone. Deep learning presents an alternative modeling approach that only requires a time history of system inputs and system states, which can be easily measured or estimated. However, fully relying on empirical or learned models involves collecting large amounts of representative data from a soft robot in order to model the complex state space-a task which may not be feasible in many situations. Furthermore, the exclusive use of empirical models for model-based control can be dangerous if the model does not generalize well. To address these challenges, we propose a hybrid modeling approach that combines machine learning methods with an existing first-principles model in order to improve overall performance for a sampling-based non-linear model predictive controller. We validate this approach on a soft robot platform and demonstrate that performance improves by 52% on average when employing the combined model.

Model Reference Predictive Adaptive Control for Large-Scale Soft Robots.

Past work has shown model predictive control (MPC) to be an effective strategy for controlling continuum joint soft robots using basic lumped-parameter models. However, the inaccuracies of these models often mean that an integral control scheme must be combined with MPC. In this paper we present a novel dynamic model formulation for continuum joint soft robots that is more accurate than previous models yet remains tractable for fast MPC. This model is based on a piecewise constant curvature (PCC) assumption and a relatively new kinematic representation that allows for computationally efficient state prediction. However, due to the difficulty in determining model parameters (e.g., inertias, damping, and spring effects) as well as effects common in continuum joint soft robots (hysteresis, complex pressure dynamics, etc.), we submit that regardless of the model selected, most model-based controllers of continuum joint soft robots would benefit from online model adaptation. Therefore, in this paper we also present a form of adaptive model predictive control based on model reference adaptive control (MRAC). We show that like MRAC, model reference predictive adaptive control (MRPAC) is able to compensate for "parameter mismatch" such as unknown inertia values. Our experiments also show that like MPC, MRPAC is robust to "structure mismatch" such as unmodeled disturbance forces not represented in the form of the adaptive regressor model. Experiments in simulation and hardware show that MRPAC outperforms individual MPC and MRAC.

Model-Based Control of Soft Actuators Using Learned Non-linear Discrete-Time Models.

Soft robots have the potential to significantly change the way that robots interact with the environment and with humans. However, accurately modeling soft robot and soft actuator dynamics in order to perform model-based control can be extremely difficult. Deep neural networks are a powerful tool for modeling systems with complex dynamics such as the pneumatic, continuum joint, six degree-of-freedom robot shown in this paper. Unfortunately it is also difficult to apply standard model-based control techniques using a neural net. In this work, we show that the gradients used within a neural net to relate system states and inputs to outputs can be used to formulate a linearized discrete state space representation of the system. Using the state space representation, model predictive control (MPC) was developed with a six degree of freedom pneumatic robot with compliant plastic joints and rigid links. Using this neural net model, we were able to achieve an average steady state error across all joints of approximately 1 and 2° with and without integral control respectively. We also implemented a first-principles based model for MPC and the learned model performed better in terms of steady state error, rise time, and overshoot. Overall, our results show the potential of combining empirical modeling approaches with model-based control for soft robots and soft actuators.

Whole-arm tactile sensing for beneficial and acceptable contact during robotic assistance.

Many assistive tasks involve manipulation near the care-receiver's body, including self-care tasks such as dressing, feeding, and personal hygiene. A robot can provide assistance with these tasks by moving its end effector to poses near the care-receiver's body. However, perceiving and maneuvering around the care-receiver's body can be challenging due to a variety of issues, including convoluted geometry, compliant materials, body motion, hidden surfaces, and the object upon which the body is resting (e.g., a wheelchair or bed). Using geometric simulations, we first show that an assistive robot can achieve a much larger percentage of end-effector poses near the care-receiver's body if its arm is allowed to make contact. Second, we present a novel system with a custom controller and whole-arm tactile sensor array that enables a Willow Garage PR2 to regulate contact forces across its entire arm while moving its end effector to a commanded pose. We then describe tests with two people with motor impairments, one of whom used the system to grasp and pull a blanket over himself and to grab a cloth and wipe his face, all while in bed at his home. Finally, we describe a study with eight able-bodied users in which they used the system to place objects near their bodies. On average, users perceived the system to be safe and comfortable, even though substantial contact occurred between the robot's arm and the user's body.