This Small Business Innovation Research (SBIR) Phase I project addresses the challenges of a decade's long unresolved technological barrier preventing a revolutionary increase in the performance of robotic systems. Enhanced productivity sought in many applications requires higher cycle rates placing great demands on robot kinematics, actuators and control systems. Further increase in operating speed must resolve dynamic challenges intrinsic in directly coupling servo actuators and robot linkages. When compared with existing electromechanical servo actuators - which route power through complex mechanical transmissions - direct-drive actuation of robot linkages enables simple mechanics and rapid motion, but does not provide dynamic isolation between the actuator and robot system. Consequently, direct-drive servo actuators are sensitive to variations in plant parameters, unknown disturbances, and un-modeled dynamics. This R&D investigates the feasibility of an innovative, direct-drive pneumatic robot actuator coupled with advanced control algorithms which rapidly accommodate dynamic system variations. Effectiveness of a new control strategy that is model insensitive, resolving unknown disturbances, un-modeled dynamics, and unknown system parameters, will be researched and developed. Success of this project will provide for a parallel delta robot that is significantly faster, more precise, possesses greater load capacity, and is substantially more affordable than contemporary delta robot systems.
The broader impact/commercial potential of this project involves engineering research conducted to enhance understanding of the dynamic interaction between direct-drive servo actuators and robotic mechanisms, and further to enhance the effectiveness and understanding of a novel control strategy which provides for an advantageous coupling between them, heretofore not practically feasible. This has the potential of introducing transformative change in the robotics industry, and to industrial automation in general. Furthermore, much of the controls knowledge gained from this research can be extended to systems employing AC linear motors, and to electromechanical servos with mismatched inertia ratios. Two market segments will be targeted: robotics and general motion control, both estimated at $7 billion. If software, peripherals and systems engineering are included, the robotics market is estimated at $19 billion. Parallel delta robots are estimated at 25% of the robotics market. The robotics industry significantly supports the national economy with applications ranging from manufacturing and food processing, to medical advances such as remotely controlled surgery, and to national defense. Well paying new hi-tech jobs are created in engineering and technical services. This research will develop revolutionary new robotic applications, educational STEM opportunities, enhanced scientific and technological understanding, making the U.S. more competitive globally.
The Phase I Project "A Pneumatically Actuated Robot System" produced three significant findings: F1) A controls technique known as Active Disturbance Rejection Control (ADRC) was considered for its preliminary promise to deliver robust performance. However, ADRC was found to be poorly suited for industrial robotic systems with substantial load variation and unknown dynamics. Fundamentally, the ADRC method creates a nested state feedback loop - which is an effective technique to normalize low frequency performance, but one that is common to the art, sensitive to electronic noise, and ultimately dependent upon a plant model. F2) A new method to predict derivatives of position has been designed, simulated, and successfully coded into the Sunstream controller firmware. Performance tests of the pneumatic actuator with the updated method showed significant improvements in positioning accuracy, and more intuitive management of servo generated noise. Reproducible positioning accuracy competitive with contemporary electric servo systems is achieved - even with high accelerations and high mass vertical loads. F3) In precision servo systems, limit cycles, or small amplitude oscillations, are often induced by friction in sliding mechanisms. A new algorithm to suppress limit cycles was simulated and embedded in real-time code. This algorithm was shown in testing to effectively suppress limit-cycles without measurably deteriorating accuracy.
