Twisted-coiled polymer actuators (TCA) have many interesting properties that show potentials for making high performance bionic devices. This letter presents a long short term memory (LSTM) model to predict the behavior of an antagonistic joint driven by hybrid TCA bundles made from Spandex and nylon fibers. By using automatic differentiation, which can be done with PyTorch, a linearized discrete state space model can be formulated and then be utilized in model predictive control (MPC). The experimental results show that by combining LSTM and MPC in modeling and control of the TCA, both high prediction performance and control performance can be achieved. It is verified that using the LSTM model, the joint angle and actuator temperatures can be predicted accurately with an avarage steady state error less than 0.1 deg and 0.2 Cel. deg., respectively. The MPC control results show the controller's ability to reach the desired joint angle with an avarage steady state error of 0.21 degrees in set-point regulation and track a sinusoidal waveform at composite frequencies of 0.1 Hz to 0.15 Hz with the steady-state error less than 0.38 degrees while changing joint stiffness.
The need for combined task and motion planning (CTAMP) in robotics is well known as robotic technologies become more mature. The goal of CTAMP is to determine a proper sequence of a robot’s actions based on symbolic and geometric reasoning. Because of the fundamental difference in symbolic and geometric reasoning, a CTAMP system often requires an interface module between the two reasoning modules. We propose a CTAMP system in which a symbolic action sequence is generated in task planning, and each action is verified geometrically in motion planning using the off-the-shelf planners and reasoners. The approach is that a set of action models is defined with PDDL in the interface module (action library) and the required information to each planner is automatically provided by the interface module. The proposed method was successfully implemented in three simulated experiments that involve manipulation tasks. According to our findings, the proposed method is effective in responding to changes in the environment and uncertainty with errors in recognition of the environment and the robot motion control.
12 March 2021 in International Journal of Precision Engineering and Manufacturing
Twisted-coiled polymer actuators (TCAs) have been known for many interesting properties such as inherent compliance, high stroke, high power density. However, the current TCAs normally come along with traditional sensing methods such as encoders attached to the link, which limits the application of the soft actuator. This letter presents a biologically inspired soft robotic muscle from TCAs with embedded proprioceptive position feedback. The embedded elongation sensor fabricated from EcoFlex and liquid metal has little hysteresis and is soft, which can help detect the muscle’s displacement. Moreover, the position of the muscle module is controlled by applying an adaptive backstepping sliding mode control algorithm which can provide finite-time convergence, high tracking performance with no singularity, and less chattering. It is verified from the experimental results that the muscle can be controlled with an average steady state error of 0.15 mm and follow a sinusoidal waveform with composite frequencies of 0.01 Hz and 0.03 Hz using natural cooling. The proposed controller has better performance over computed torque method control and adaptive sliding mode control in terms of maximum error, mean steady state error, rising time, peak time, and overshoot. A robotic hand application driven by the proposed artificial module was also demonstrated.
Autonomous street sweeping is an interesting robotic application because such street cleaning requires tedious and exhaustive human efforts of manual sweeping with brooms, blowing tools, or driving sweeping vehicles within a large area or along a long extension of roads on a regular basis. Toward autonomous street cleaning, we present our recent development of an autonomous robotic street sweeper. The developed system is aimed at autonomous cleaning with low cost hardware. In this work, a commercialized manual brush sweeper is robotized and equipped with two fish-eye cameras for environmental sensing. As an initial attempt of autonomous street sweeping, we show the cleaning performance of the robot along the curbside based on curbside recognition and tracking with top-view images.
Stiffness adjustability is one of the key characteristics that help humans achieve safe and reliable motions. This paper aims to realize the simultaneous position-stiffness control of an antagonistic joint driven by a type of super-coiled polymer (SCP) artificial muscles, made from a combination of Spandex and nylon fibers. A nonlinear model that can exhibit the variable stiffness characteristics of the actuator is first proposed. The model, whose parameters are identified and verified experimentally, is suitable for estimating the actuator's stiffness as a function of the length and temperature. Based on a linearized model of the antagonistic joint system, a model predictive control (MPC) is applied to control joint angle and stiffness simultaneously. By controlling both actuators' temperatures, simultaneous position-stiffness control is realized. Our contribution also includes enhancement of MPC control by incorporating time delay estimation to estimate model variations and external disturbances. The control performance is verified with both step command and sinusoidal reference with composite frequencies of 0.02Hz to 0.1Hz. Experimental results show that the system can achieve high accuracy of joint position control performance with the maximum position error of 0.42 degrees while varying joint stiffness 46.5%.