A Novel Approach to Time Series Forecasting Using Model-Free Adaptive Control Framework

Sensor data predictions could significantly improve the accuracy and effectiveness of modern control systems; however, existing machine learning and advanced statistical techniques to forecast time series data require significant computational resources which is not ideal for real-time applications. In this paper, we propose a novel forecasting technique called Compact Form Dynamic Linearization Model-Free Prediction (CFDL-MFP) which is derived from the existing model-free adaptive control framework. This approach enables near real-time forecasts of seconds-worth of time-series data due to its basis as an optimal control problem. The performance of the CFDL-MFP algorithm was evaluated using four real datasets including: force sensor readings from surgical needle, ECG measurements for heart rate, and atmospheric temperature and Nile water level recordings. On average, the forecast accuracy of CFDL-MFP was 28% better than the benchmark Autoregressive Integrated Moving Average (ARIMA) algorithm. The maximum computation time of CFDL-MFP was 49.1ms which was 170 times faster than ARIMA. Forecasts were best for deterministic data patterns, such as the ECG data, with a minimum average root mean squared error of (0.2±0.2).

Design and Analysis of a New Mechanism for a Snake Like Robot

In this paper, a novel snake like robot design is presented and analyzed. The structure described desires to obtain a robot that is most like a snake found in nature. This is achieved with the combination of both rigid and soft link structures by implementing a 3D printed rigid link and a soft cast silicone skin. The proposed structure serves to have a few mechanical improvements while maintaining the positives of previous designs. The implementation of the silicone skin presents the opportunity to use synthetic scales and directional friction. The design modifications of this novel design are analyzed on the fronts of the kinematics and minimizing power loss. Minimization of power loss is done through a numerical minimization of three separate parameters with the smallest positive power loss being used. This results in the minimal power loss per unit distance. This research found that the novel structure presented can be effectively described and modeled, such that they could be applied to a constructed model.

Development of Wire Actuated Monolithic Soft Gripper Positioned by Robot Manipulator

This paper presents the design and development of a two fingered, monolithically designed compliant gripper mounted on a two-link robot. Rigid grippers traditionally designed by rigid links and joints might have low precision due to friction and backlash. The proposed gripper is designed as a single piece compliant mechanism consisted of several flexible links and actuated by wire through a servo motor. The gripper is attached to a two-link arm robot driven by three step motors. An additional servo motor can also rotate the base of the robot. While the robot is 3D printed using polylactic acid (PLA), the gripper is 3D printed in thermoplasticpolyurethane (TPU). Two force sensors are attached to the right and left ends of the gripper to measure grasping force. Experimental testing for grasping various objects having different sizes, shapes and weights is carried out to verify the robust performance of the proposed design. Through the experimentation, it’s been noted that the compliant gripper can successfully lift up objects at a maximum mass of 200 g and have a better performance if the objects’width is closer to the width of the gripper. The presented mechanism can be utilized as a service robot for elderly people to assist them pick and place objects or lift objects if equipped with necessary sensors.

Cutting the Deployment Costs of Physics-Based MPC in Buildings by Simulation-Based Imitation Learning

It has been shown that model predictive control (MPC) is a promising solution for energy-efficient building operations. However, the deployment of MPC in a large portion of the building stock has not been possible partially because of high installation costs. Every building is unique and requires a tailored MPC solution. The best performing solutions are often based on physics-based modeling, which is, however, computationally expensive and requires dedicated software. A promising direction that tackles this problem is to train a neural network-based optimal control policy to imitate the behavior of physics-based MPC from the simulation data generated offline. The neural networks give control actions that closely approximate those produced by physics-based MPC, but with a fraction of the computational and memory requirements and without the need for licensed software. The main advantage of the proposed approach stems from simple evaluation at execution time, leading to low computational foot-prints and easy deployment on embedded HW platforms. In the case study, we present the energy savings potential of physics-based MPC applied to an office building in Belgium. We demonstrate how neural network approximators can be used to cut the implementation and maintenance costs of MPC deployment without compromising performance. We also critically assess the presented approach by pointing out the remaining challenges and open research questions.

Parametric Modeling and Optimal Control of a Combined Heating and Power System With Energy Storage

Cogeneration is a well-known and cost effective solution for generating power and heat within the same plant, leading to improved overall efficiency and reduced generation cost. Combined heating and power systems can facilitate the penetration of renewable energy sources in medium size applications through the integration of electric and thermal energy storage units. Due to the complexity of the plant as well as significantly variability in power demand and generation, the design and operation of such systems requires a systematic co-optimization of plant and controller for guaranteeing near optimal performance. In this scenario, this paper presents a physics-based parametric modeling approach for the characterization of the main components of a 1MW combined heating and power system that includes renewable sources, electric and thermal storage devices. To demonstrate the model flexibility and potential benefits achieved by an optimal sizing, the system energy management is optimized using Dynamic Programming. The operational costs for different configurations are compared showing that an optimization of the energy management strategy in conjunction with an improved system sizing lead to more than 6% of reduction in the operational cost.

The Swirling Pendulum: Conceptualization, Modeling, Equilibria and Control Synthesis

This paper introduces the swirling pendulum, a two-link, two degree-of-freedom mechanism which is under-actuated and has an unusual non-planar coupling with axis of rotation of the two links being perpendicular to each other. The swirling pendulum mechanism, while being simple to mathematically represent and easy to physically construct, exhibits several properties like loss of inertial coupling, loss of relative degree, multiple stable and unstable equilibrium points. These properties are unique as well as interesting from dynamics and controls point of view which make the swirling pendulum an excellent test-bed for testing various ideas in control and demonstrating several notions associated with systems and control theory. In this paper, we discuss the modeling of the swirling pendulum mechanism based on Lagrange’s equation along with an analysis related to equilibrium points and their stability. We also present simulation results for regulatory as well as tracking control tasks through simulations on a non-linear model using control methods like LQR, lead compensator and system inversion-based control to demonstrate the utility of the proposed mechanism in the area of systems, control and dynamics. Furthermore, we also discuss experimental results for controls applied on a real-time hardware setup.

Minimum Safety Distances for Emergency Braking Maneuvers in Car-Following Applications

Platooning/car following has been considered as a promising approach for improving vehicle efficiency due to the reduction of aerodynamic force when closely following a pilot vehicle. However, safety is a major concern in the close car platooning/following. This paper investigates the minimum inter-vehicle distances required for a passenger vehicle to safely travel behind a heavy-duty truck with three different types of emergency maneuvers. The three emergency maneuvers considered are braking only, steering only, and braking then steering, where steering refers to a single lane change maneuver. Numerical analysis is conducted for deriving the clearance space in the braking only scenario. In addition, simulations are conducted in MATLAB/Simulink, using a bicycle model for the vehicle dynamics, to examine the minimum safe following distance for the other two scenarios. The simulation results show that, for initial vehicle speeds greater than 8 m/s, a lane change maneuver requires the shortest safety distance. Braking followed by lane changing usually requires the largest minimum safety distance.

Constrained Control of Input Delayed Systems With Partially Compensated Input Delays

Control Barrier Functions (CBFs) have become popular for enforcing — via barrier constraints — the safe operation of nonlinear systems within an admissible set. For systems with input delay(s) of the same length, constrained control has been achieved by combining a CBF for the delay free system with a state predictor that compensates the single input delay. Recently, this approach was extended to multi input systems with input delays of different lengths. One limitation of this extension is that barrier constraint adherence can only be guaranteed after the longest input delay has been compensated and all input channels become available for control. In this paper, we consider the problem of enforcing constraint adherence when only a subset of input delays have been compensated. In particular, we propose a new barrier constraint formulation that ensures that when possible, a subset of input channels with shorter delays will be utilized for keeping the system in the admissible set even before longer input delays have been compensated. We include a numerical example to demonstrate the effectiveness of the proposed approach.

PD Controller With Self Adaptive Gains for Quadrotor Waypoint Navigation

Proportional-Derivative (PD) controllers are commonly used in quadrotors due to their simple structure. Tuning of the gains of the PD controller is often cumbersome due to strong coupling of the dynamics between three linear and three angular degrees of freedom. This paper presents a novel method of auto adjusting the proportional and derivative gains of the quadrotor without the use of any stable reference model (unlike model reference adaptive control). The gains are automatically adjusted throughout the flight based on just the state errors. Lyapunov stability analysis and adaptive gain law is used to formulate the control algorithm to achieve way point navigation. It is shown that our proposed controller achieves effective way point navigation even when started off from random gain values.

On a Gamified Brain-Computer Interface for Cognitive Training of Spatial Working Memory

In the United States, there are a large number of people suffering from memory and attention deficit problems. For example, patients with attention-deficit hyperactivity disorder (ADHD) and dementia have difficulties in performing activities of daily living and have a low quality of life. Currently, there exist no effective treatment for these memory and attention issues in specific cognitive impairments. In this paper, we developed a gamified platform of brain-computer interface (BCI) for cognitive training, which can engage users in the training and provide users qualitative and quantitative feedback for their training of spatial working memory. The user is able to control the movement of a drone using motor imager, which is imagined movement of body part. Sensorimotor rhythms of the user are calculated using the user’s EEG to drive the movement of the drone. Twenty normal healthy subjects were recruited to test the user experience. Our system showed the capability of engaging users, good robustness, user acceptability and usability. Therefore, we think our platform might be an alternative to provide more accessible, engaging, and effective cognitive training for people with memory and attention problems. In future, we will test the usability and effectiveness of the system for cognitive training in patients with ADHD and dementia.