This article considers collision avoidance (COLAV) for both static and moving obstacles using the branching-course model predictive control (BC-MPC) algorithm, which is designed for use by autonomous surface vehicles (ASVs). The BC-MPC algorithm originally only considered COLAV of moving obstacles, so in order to make the algorithm also be able to avoid static obstacles, we introduce an extra term in the objective function based on an occupancy grid. In addition, other improvements are made to the algorithm resulting in trajectories with less wobbling. The modified algorithm is verified through full-scale experiments in the Trondheimsfjord in Norway with both virtual static obstacles and a physical moving obstacle. A radar-based tracking system is used to detect and track the moving obstacle, which enables the algorithm to avoid obstacles without depending on vessel-to-vessel communication. The experiments show that the algorithm is able to simultaneously avoid both static and moving obstacles, while providing clear and readily observable maneuvers. The BC-MPC algorithm is compliant with rules 8, 13 and 17 of the the International Regulations for Preventing Collisions at Sea (COLREGs), and favors maneuvers following rules 14 and 15.
This paper presents a three-layered hybrid collision avoidance (COLAV) system for autonomous surface vehicles, compliant with rules 8 and 13-17 of the International Regulations for Preventing Collisions at Sea (COLREGs). The COLAV system consists of a high-level planner producing an energy-optimized trajectory, a model predictive control based mid-level COLAV algorithm considering moving obstacles and the COLREGs, and the branching-course model predictive control algorithm for short-term COLAV handling emergency situations in accordance with the COLREGs. Previously developed algorithms by the authors are used for the high-level planner and short-term COLAV, while we in this paper further develop the mid-level algorithm to make it comply with COLREGs rules 13-17. This includes developing a state machine for classifying obstacle vessels using a combination of the geometrical situation, the distance and time to the closest point of approach (CPA) and a new CPA-like measure. The performance of the hybrid COLAV system is tested through numerical simulations for three scenarios representing a range of different challenges, including multi-obstacle situations with multiple simultaneously active COLREGs rules, and also obstacles ignoring the COLREGs. The COLAV system avoids collision in all the scenarios, and follows the energy-optimized trajectory when the obstacles do not interfere with it.
We formulate the optimal energy arbitrage problem for a piecewise linear cost function for energy storage devices using linear programming (LP). The LP formulation is based on the equivalent minimization of the epigraph. This formulation considers ramping and capacity constraints, charging and discharging efficiency losses of the storage, inelastic consumer load and local renewable generation in presence of net-metering which facilitates selling of energy to the grid and incentivizes consumers to install renewable generation and energy storage. We consider the case where the consumer loads, electricity prices, and renewable generations at different instances are uncertain. These uncertain quantities are predicted using an Auto-Regressive Moving Average (ARMA) model and used in a model predictive control (MPC) framework to obtain the arbitrage decision at each instance. In numerical results we present the sensitivity analysis of storage performing arbitrage with varying ramping batteries and different ratio of selling and buying price of electricity.
This paper proposes a methodology of model predictive control for alleviating shallow drowsiness of office workers and thus improving their productivity. The methodology is based on dynamically scheduling setting values for air conditioning and lighting to minimize drowsiness level of office workers on the basis of a prediction model that represents the relation between future drowsiness level and combination of indoor temperature and ambient illuminance. The prediction model can be identified by utilizing state-of-the-art drowsiness estimation method. The proposed methodology was evaluated in regard to a real routine task (performed by six subjects over five workdays), and the evaluation results demonstrate that the proposed methodology improved the processing speed of the task by 8.3% without degrading comfort of the workers.
Collision avoidance is a critical task in many applications, such as ADAS (advanced driver-assistance systems), industrial automation and robotics. In an industrial automation setting, certain areas should be off limits to an automated vehicle for protection of people and high-valued assets. These areas can be quarantined by mapping (e.g., GPS) or via beacons that delineate a no-entry area. We propose a delineation method where the industrial vehicle utilizes a LiDAR {(Light Detection and Ranging)} and a single color camera to detect passive beacons and model-predictive control to stop the vehicle from entering a restricted space. The beacons are standard orange traffic cones with a highly reflective vertical pole attached. The LiDAR can readily detect these beacons, but suffers from false positives due to other reflective surfaces such as worker safety vests. Herein, we put forth a method for reducing false positive detection from the LiDAR by projecting the beacons in the camera imagery via a deep learning method and validating the detection using a neural network-learned projection from the camera to the LiDAR space. Experimental data collected at Mississippi State University's Center for Advanced Vehicular Systems (CAVS) shows the effectiveness of the proposed system in keeping the true detection while mitigating false positives.
We introduce a microscopic model of interacting financial agents, where each agent is characterized by two portfolios; money invested in bonds and money invested in stocks. Furthermore, each agent is faced with an optimization problem in order to determine the optimal asset allocation. The stock price evolution is driven by the aggregated investment decision of all agents. In fact, we are faced with a differential game since all agents aim to invest optimal. Mathematically such a problem is ill posed and we introduce the concept of Nash equilibrium solutions to ensure the existence of a solution. Especially, we denote an agent who solves this Nash equilibrium exactly a rational agent. As next step we use model predictive control to approximate the control problem. This enables us to derive a precise mathematical characterization of the degree of rationality of a financial agent. This is a novel concept in portfolio optimization and can be regarded as a general approach. In a second step we consider the case of a fully myopic agent, where we can solve the optimal investment decision of investors analytically. We select the running cost to be the expected missed revenue of an agent and we assume quadratic transaction costs. More precisely the expected revenues are determined by a combination of a fundamentalist or chartist strategy. Then we derive the mean field limit of the microscopic model in order to obtain a macroscopic portfolio model. The novelty in comparison to existent macroeconomic models in literature is that our model is derived from microeconomic dynamics. The resulting portfolio model is a three dimensional ODE system which enables us to derive analytical results. Simulations reveal, that our model is able to replicate the most prominent features of financial markets, namely booms and crashes.
We present a novel control policy, called Predictive Network Control (PNC) to control wireless communication networks (on packet level), based on paradigms of Model Predictive Control (MPC). In contrast to common myopic policies, who use one step ahead prediction, PNC predicts the future behavior of the system for an extended horizon, thus facilitating performance gains. We define an advanced system model in which we use a Markov chain in combination with a Bernoulli trial to model the stochastic components of the network. Furthermore we introduce the algorithm and present two detailed simulation examples, which show general improved performance and a gain in stability region compared to the standard policy.
In control applications there is often a compromise that needs to be made with regards to the complexity and performance of the controller and the computational resources that are available. For instance, the typical hardware platform in embedded control applications is a microcontroller with limited memory and processing power, and for battery powered applications the control system can account for a significant portion of the energy consumption. We propose a controller architecture in which the computational cost is explicitly optimized along with the control objective. This is achieved by a three-part architecture where a high-level, computationally expensive controller generates plans, which a computationally simpler controller executes by compensating for prediction errors, while a recomputation policy decides when the plan should be recomputed. In this paper, we employ model predictive control (MPC) as the high-level plan-generating controller, a linear state feedback controller as the simpler compensating controller, and reinforcement learning (RL) to learn the recomputation policy. Simulation results for two examples showcase the architecture's ability to improve upon the MPC approach and find reasonable compromises weighing the performance on the control objective and the computational resources expended.
Model Predictive Control in buildings can significantly reduce their energy consumption. The cost and effort necessary for creating and maintaining first principle models for buildings make data-driven modelling an attractive alternative in this domain. In MPC the models form the basis for an optimization problem whose solution provides the control signals to be applied to the system. The fact that this optimization problem has to be solved repeatedly in real-time implies restrictions on the learning architectures that can be used. Here, we adapt Input Convex Neural Networks that are generally only convex for one-step predictions, for use in building MPC. We introduce additional constraints to their structure and weights to achieve a convex input-output relationship for multistep ahead predictions. We assess the consequences of the additional constraints for the model accuracy and test the models in a real-life MPC experiment in an apartment in Switzerland. In two five-day cooling experiments, MPC with Input Convex Neural Networks is able to keep room temperatures within comfort constraints while minimizing cooling energy consumption.
This paper presents DeepKoCo, a novel model-based agent that learns a latent Koopman representation from images. This representation allows DeepKoCo to plan efficiently using linear control methods, such as linear model predictive control. Compared to traditional agents, DeepKoCo learns task-relevant dynamics, thanks to the use of a tailored lossy autoencoder network that allows DeepKoCo to learn latent dynamics that reconstruct and predict only observed costs, rather than all observed dynamics. As our results show, DeepKoCo achieves similar final performance as traditional model-free methods on complex control tasks while being considerably more robust to distractor dynamics, making the proposed agent more amenable for real-life applications.
This paper investigates the problem of designing data-driven stochastic Model Predictive Control (MPC) for linear time-invariant systems under additive stochastic disturbance, whose probability distribution is unknown but can be partially inferred from data. We propose a novel online learning based risk-averse stochastic MPC framework in which Conditional Value-at-Risk (CVaR) constraints on system states are required to hold for a family of distributions called an ambiguity set. The ambiguity set is constructed from disturbance data by leveraging a Dirichlet process mixture model that is self-adaptive to the underlying data structure and complexity. Specifically, the structural property of multimodality is exploit-ed, so that the first- and second-order moment information of each mixture component is incorporated into the ambiguity set. A novel constraint tightening strategy is then developed based on an equivalent reformulation of distributionally ro-bust CVaR constraints over the proposed ambiguity set. As more data are gathered during the runtime of the controller, the ambiguity set is updated online using real-time disturbance data, which enables the risk-averse stochastic MPC to cope with time-varying disturbance distributions. The online variational inference algorithm employed does not require all collected data be learned from scratch, and therefore the proposed MPC is endowed with the guaranteed computational complexity of online learning. The guarantees on recursive feasibility and closed-loop stability of the proposed MPC are established via a safe update scheme. Numerical examples are used to illustrate the effectiveness and advantages of the proposed MPC.
Modern vehicles communicate data to and from sensors, actuators, and electronic control units (ECUs) using Controller Area Network (CAN) bus, which operates on differential signaling. An autonomous ECU responsible for the execution of decision commands to an autonomous vehicle is developed by assimilating the information from the CAN bus. The conventional way of parsing the decision commands is motion planning, which uses a path tracking algorithm to evaluate the decision commands. This study focuses on designing a robust controller using behavioral cloning and motion planning of autonomous vehicle using a deep learning framework. In the first part of this study, we explore the pipeline of parsing decision commands from the path tracking algorithm to the controller and proposed a neural network-based controller ($N^2C$) using behavioral cloning. The proposed network predicts throttle, brake, and torque when trained with the manual driving data acquired from the CAN bus. The efficacy of the proposed method is demonstrated by comparing the accuracy with the Proportional-Derivative-Integral (PID) controller in conjunction with the path tracking algorithm (pure pursuit and model predictive control based path follower). The second part of this study complements $N^2C$, in which an end-to-end neural network for predicting the speed and steering angle is proposed with image data as an input. The performance of the proposed frameworks are evaluated in real-time and on the Udacity dataset, showing better metric scores in the former and reliable prediction in the later case when compared with the state-of-the-art methods.
We consider the problem of designing an algorithm to allow a car to autonomously merge on to a highway from an on-ramp. Two broad classes of techniques have been proposed to solve motion planning problems in autonomous driving: Model Predictive Control (MPC) and Reinforcement Learning (RL). In this paper, we first establish the strengths and weaknesses of state-of-the-art MPC and RL-based techniques through simulations. We show that the performance of the RL agent is worse than that of the MPC solution from the perspective of safety and robustness to out-of-distribution traffic patterns, i.e., traffic patterns which were not seen by the RL agent during training. On the other hand, the performance of the RL agent is better than that of the MPC solution when it comes to efficiency and passenger comfort. We subsequently present an algorithm which blends the model-free RL agent with the MPC solution and show that it provides better trade-offs between all metrics -- passenger comfort, efficiency, crash rate and robustness.
Humans quickly solve tasks in novel systems with complex dynamics, without requiring much interaction. While deep reinforcement learning algorithms have achieved tremendous success in many complex tasks, these algorithms need a large number of samples to learn meaningful policies. In this paper, we present a task for navigating a marble to the center of a circular maze. While this system is very intuitive and easy for humans to solve, it can be very difficult and inefficient for standard reinforcement learning algorithms to learn meaningful policies. We present a model that learns to move a marble in the complex environment within minutes of interacting with the real system. Learning consists of initializing a physics engine with parameters estimated using data from the real system. The error in the physics engine is then corrected using Gaussian process regression, which is used to model the residual between real observations and physics engine simulations. The physics engine augmented with the residual model is then used to control the marble in the maze environment using a model-predictive feedback over a receding horizon. To the best of our knowledge, this is the first time that a hybrid model consisting of a full physics engine along with a statistical function approximator has been used to control a complex physical system in real-time using nonlinear model-predictive control (NMPC).
Decision making under uncertainty is critical to real-world, autonomous systems. Model Predictive Control (MPC) methods have demonstrated favorable performance in practice, but remain limited when dealing with complex probability distributions. In this paper, we propose a generalization of MPC that represents a multitude of solutions as posterior distributions. By casting MPC as a Bayesian inference problem, we employ variational methods for posterior computation, naturally encoding the complexity and multi-modality of the decision making problem. We present a Stein variational gradient descent method to estimate the posterior directly over control parameters, given a cost function and observed state trajectories. We show that this framework leads to successful planning in challenging, non-convex optimal control problems.
Reinforcement learning (RL) is a control approach that can handle nonlinear stochastic optimal control problems. However, despite the promise exhibited, RL has yet to see marked translation to industrial practice primarily due to its inability to satisfy state constraints. In this work we aim to address this challenge. We propose an 'oracle'-assisted constrained Q-learning algorithm that guarantees the satisfaction of joint chance constraints with a high probability, which is crucial for safety critical tasks. To achieve this, constraint tightening (backoffs) are introduced and adjusted using Broyden's method, hence making them self-tuned. This results in a general methodology that can be imbued into approximate dynamic programming-based algorithms to ensure constraint satisfaction with high probability. Finally, we present case studies that analyze the performance of the proposed approach and compare this algorithm with model predictive control (MPC). The favorable performance of this algorithm signifies a step toward the incorporation of RL into real world optimization and control of engineering systems, where constraints are essential in ensuring safety.
We present Model-Predictive Interaction Primitives -- a robot learning framework for assistive motion in human-machine collaboration tasks which explicitly accounts for biomechanical impact on the human musculoskeletal system. First, we extend Interaction Primitives to enable predictive biomechanics: the prediction of future biomechanical states of a human partner conditioned on current observations and intended robot control signals. In turn, we leverage this capability within a model-predictive control strategy to identify the future ergonomic and biomechanical ramifications of potential robot actions. Optimal control trajectories are selected so as to minimize future physical impact on the human musculoskeletal system. We empirically demonstrate that our approach minimizes knee or muscle forces via generated control actions selected according to biomechanical cost functions. Experiments are performed in synthetic and real-world experiments involving powered prosthetic devices.
Motion planning in environments with multiple agents is critical to many important autonomous applications such as autonomous vehicles and assistive robots. This paper considers the problem of motion planning, where the controlled agent shares the environment with multiple uncontrolled agents. First, a predictive model of the uncontrolled agents is trained to predict all possible trajectories within a short horizon based on the scenario. The prediction is then fed to a motion planning module based on model predictive control. We proved generalization bound for the predictive model using three different methods, post-bloating, support vector machine (SVM), and conformal analysis, all capable of generating stochastic guarantees of the correctness of the predictor. The proposed approach is demonstrated in simulation in a scenario emulating autonomous highway driving.
3D scene representation for robot manipulation should capture three key object properties: permanency -- objects that become occluded over time continue to exist; amodal completeness -- objects have 3D occupancy, even if only partial observations are available; spatiotemporal continuity -- the movement of each object is continuous over space and time. In this paper, we introduce 3D Dynamic Scene Representation (DSR), a 3D volumetric scene representation that simultaneously discovers, tracks, reconstructs objects, and predicts their dynamics while capturing all three properties. We further propose DSR-Net, which learns to aggregate visual observations over multiple interactions to gradually build and refine DSR. Our model achieves state-of-the-art performance in modeling 3D scene dynamics with DSR on both simulated and real data. Combined with model predictive control, DSR-Net enables accurate planning in downstream robotic manipulation tasks such as planar pushing. Video is available at https://youtu.be/GQjYG3nQJ80.
In this paper, we develop a model to plan energy-efficient speed trajectories of electric trucks in real-time by taking into account the information of topography and traffic ahead of the vehicle. In this real time control model, a novel state-space model is first developed to capture vehicle speed, acceleration, and state of charge. We then formulate an energy minimization problem and solve it by an alternating direction method of multipliers (ADMM) method that exploits the structure of the problem. A model predictive control framework is then employed to deal with topographic and traffic uncertainties in real-time. An empirical study is conducted on the performance of the proposed eco-driving algorithm and its impact on battery degradation. The experimental results show that the energy consumption by using the developed method is reduced by up to 5.05%, and the battery life extended by as high as 35.35% compared to benchmarking solutions.