Sangram Redkar

Limit Cycles to Enhance Human Performance Based on Phase Oscillators

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. This “energy pumping” at resonance can reduce the metabolic cost of performing cyclic tasks. Many human tasks such as walking, running, and hopping are repeating or cyclic tasks where assistance is needed at a repeating rate at the correct time. By utilizing resonant energy pumping, a tiny amount of energy is added at an appropriate time that results in an amplified response. However, when the system dynamics is varying or uncertain, resonant boundaries are not clearly defined. We have developed a method to add energy at resonance so the system attains the limit cycle based on a phase oscillator. The oscillator is robust to disturbances and initial conditions and allows our robots to enhance running, reduce metabolic cost, and increase hop height. These methods are general and can be used in other areas such as energy harvesting.

Fuzzy logic based sensor fusion for accurate tracking

Accuracy and tracking update rates play a vital role in determining the quality of Augmented Reality(AR) and Virtual Reality(VR) applications. Applications like soldier training, gaming, simulations & virtual conferencing need a high accuracy tracking with update frequency above 20Hz for an immersible experience of reality. Current research techniques combine more than one sensor like camera, infrared, magnetometers and Inertial Measurement Units (IMU) to achieve this goal. In this paper, we develop and validate a novel algorithm for accurate positioning and tracking using inertial and vision-based sensing techniques. The inertial sensing utilizes accelerometers and gyroscopes to measure rates and accelerations in the body fixed frame and computes orientations and positions via integration. The vision-based sensing uses camera and image processing techniques to compute the position and orientation. The sensor fusion algorithm proposed in this work uses the complementary characteristics of these two independent systems to compute an accurate tracking solution and minimizes the error due to sensor noise, drift and different update rates of camera and IMU. The algorithm is computationally efficient, implemented on a low cost hardware and is capable of an update rate up to 100 Hz. The position and orientation accuracy of the sensor fusion is within 6mm & 1.5°. By using the fuzzy rule sets and adaptive filtering of data, we reduce the computational requirement less than the conventional methods (such as Kalman filtering). We have compared the accuracy of this sensor fusion algorithm with a commercial infrared tracking system. It can be noted that outcome accuracy of this COTS IMU and camera sensor fusion approach is as good as the commercial tracking system at a fraction of the cost.

HeSA, Hip Exoskeleton for Superior Assistance

A hip exoskeleton was designed that can assist hip flexion and extension. The device incorporates a motor, ball-screw, and spring in a lightweight package. The total weight including the battery is 2.95 kg. The system uses 20 W of power per leg. The system is controlled based on the phase angle of each leg and the torque is applied in synchrony with the user’s steps. The device assists walking, running, and does not interfere when going up and down stairs.

Reduced-Order Modeling of Parametrically Excited Micro-Electro-Mechanical Systems (MEMS)

Reduced-order modeling is a systematic way of constructing models with smaller number of states that can capture the “essential dynamics” of the large-scale systems, accurately. In this paper, reduced-order modeling and control techniques for parametrically excited MEMS are presented. The techniques proposed here use the Lyapunov-Floquet (L-F) transformation that makes the linear part of transformed equations time invariant. In this work, three model reduction techniques for MEMS are suggested. First method is simply an application of the well-known Guyan-like reduction method to nonlinear systems. The second technique is based on singular perturbation, where the transformed system dynamics is partitioned as fast and slow dynamics and the system of differential equations is converted into a differential algebraic (DAE) system. In the third technique, the concept of invariant manifold for time-periodic systems is used. The “time periodic invariant manifold” based technique yields “reducibility conditions”. This is an important result because it helps us to understand the various types of resonances present in the system. These resonances indicate a tight coupling between the system states, and in order to retain the dynamic characteristics, one has to preserve all these “resonant” states in the reduced-order model. Thus, if the “reducibility conditions” are satisfied, only then a nonlinear order reduction based on invariant manifold approach is possible. It is found that the invariant manifold approach yields the most accurate results followed by the nonlinear projection and linear technique. These methodologies are general, free from small parameter assumptions, and can be applied to a variety of MEM systems like resonators, sensors and filters. The reduced-order models can be used for parametric study, sensitivity analysis and/or controller design. The controller design is based on the reduced-order system. Thus, first the reduced-order model of the large-scale system is constructed that captures the essential dynamics. If a controller is designed to stabilize this reduced-order system, then it guarantees that the large-scale system is controlled. The theoretical framework to design linear and nonlinear controllers is also presented.

Human activity recognition using inertial measurement units and smart shoes

This paper presents an intelligent fuzzy inference (IFI) algorithm using inertial measurement units (IMUs) and smart shoes to recognize human activities. IFI algorithm recognizes the activities based on ground contact forces (GCFs) and the knee joint angles. The smart shoes are designed to measure GCFs exerted by the wearer. A total of four IMUs are mounted on bilateral thighs and shanks to provide acceleration and angular rate data. To calculate knee flexion extension, a calibration procedure is adopted which eliminates the need for an external camera system. Then, an extended Kalman filter (EKF) is used to estimate the relative orientations of thigh and shank segments, from which knee angle is calculated. Random forest search (RFS) technique is used as a baseline to compare with the performance of the IFI algorithm. To evaluate the performance of this algorithm, several outdoor experiments are conducted on two healthy subjects for six activities including sitting, standing, walking, going upstairs, going downstairs and jogging. The results show that the algorithm is capable of classifying six activities with higher precision and less update time compared to the baseline approach for both subject dependent and independent tests. Also, the algorithm detects transitions between all the activities smoothly such as sit-to-stand or stand-to-walk with higher precision.

Bioinspired controller based on a phase oscillator

Oscillatory behavior is important for tasks such as walking and running. We are developing methods to add energy to enhance or vary the oscillatory behavior based on the phase angle. Both the amplitude and frequency of oscillation can be modulated by adjusting the forcing function based on the sine and cosine of the phase angle. Pendulum systems are simulated using our phase controller. We show that a double-pendulum can be controlled using an oscillator.

Analysis and simulation of Wiseman hypocycloid engine

Abstract This research studies an alternative to the slider-crank mechanism for internal combustion engines, which was proposed by the Wiseman Technologies Inc. Their design involved replacing the crankshaft with a hypocycloid gear assembly. The unique hypocycloid gear arrangement allowed the piston and connecting rod to move in a straight line creating a perfect sinusoidal motion, without any side loads. In this work, the Wiseman hypocycloid engine was modeled in a commercial engine simulation software and compared to slider-crank engine of the same size. The engine’s performance was studied, while operating on diesel, ethanol, and gasoline fuel. Furthermore, a scaling analysis on the Wiseman engine prototypes was carried out to understand how the performance of the engine is affected by increasing the output power and cylinder displacement. It was found that the existing 30cc Wiseman engine produced about 7% less power at peak speeds than the slider-crank engine of the same size. These results were concurrent with the dynamometer tests performed in the past. It also produced lower torque and was about 6% less fuel efficient than the slider-crank engine. The four-stroke diesel variant of the same Wiseman engine performed better than the two-stroke gasoline version. The Wiseman engine with a contra piston (that allowed to vary the compression ratio) showed poor fuel efficiency but produced higher torque when operating on E85 fuel. It also produced about 1.4% more power than while running on gasoline. While analyzing effects of the engine size on the Wiseman hypocycloid engine prototypes, it was found that the engines performed better in terms of power, torque, fuel efficiency, and cylinder brake mean effective pressure as the displacement increased. The 30 horsepower (HP) conceptual Wiseman prototype, while operating on E85, produced the most optimum results in all aspects, and the diesel test for the same engine proved to be the most fuel efficient.

Inertial Sensing of Dummy Kinematics

Motor vehicle crashes claim over 40,000 lives and injure over two million people each year in the United States. To reduce the number of injuries and fatalities through vehicle design improvements, it is important to study occupant kinematics and related injury mechanisms during crashes. Occupant motion in crash tests is typically measured with high speed video, spatial scanning, direct field sensing, and inertial sensing. In this work, we present simulation and testing results on inertial sensing of dummy kinematics based on a novel algorithm known as Quaternion Fuzzy Logic Adaptive Signal Processing for Biomechanics (QFLASP-B). This approach uses three angular rates and three accelerations (one gyroscope-accelerometer pair about each axis) per rigid body to compute orientations (roll, pitch and yaw), positions and velocities in the inertial (fixed) reference frame. In QFLASP-B, quaternion errors and gyro biases are calculated and used in an adaptive loop to remove their effects. The Fuzzy Estimator at the core of the algorithm consists of a fuzzification process, an inference mechanism, a Rule Base and a defuzzification process. In this paper, we examine those aspects of the QFLASP-B Fuzzy Estimator critical to accurate kinematics sensing, hardware and software implementations and experimental results compared with traditional approaches.Copyright