All you need to know about model predictive control for buildings

We study the problem of finite-time constrained optimal control of unknown stochastic linear time-invariant systems, which is the key ingredient of a predictive control algorithm - albeit typically having access to a model. We propose a novel distributionally robust data-enabled predictive control (DeePC) algorithm which uses noise-corrupted input/output data to predict future trajectories and compute optimal control inputs while satisfying output chance constraints. The algorithm is based on (i) a non-parametric representation of the subspace spanning the system behaviour, where past trajectories are sorted in Page or Hankel matrices; and (ii) a distributionally robust optimization formulation which gives rise to strong probabilistic performance guarantees. We show that for certain objective functions, DeePC exhibits strong out-of-sample performance, and at the same time respects constraints with high probability. The algorithm provides an end-to-end approach to control design for unknown stochastic linear time-invariant systems. We illustrate the closed-loop performance of the DeePC in an aerial robotics case study.

/pdf/distributionally-robust-chance-constrained-data-enabled-4hth8s9zz6.pdf

Distributionally Robust Chance Constrained Data-enabled Predictive Control

We propose a new method to obtain feedback controllers of an unknown dynamical system directly from noisy input/state data. The key ingredient of our design is a new matrix S-lemma that will be proven in this paper. We provide both strict and non-strict versions of this S-lemma, that are of interest in their own right. Thereafter, we will apply these results to data-driven control. In particular, we will derive non-conservative design methods for quadratic stabilization, H_2 and H_inf control, all in terms of data-based linear matrix inequalities. In contrast to previous work, the dimensions of our decision variables are independent of the time horizon of the experiment. Our approach thus enables control design from large data sets.

From noisy data to feedback controllers: non-conservative design via a matrix S-lemma

The Department of Transport in the United Kingdom recorded 25,080 motor vehicle fatalities in 2019. This situation stresses the need for an intelligent transport system (ITS) that improves road safety and security by avoiding human errors with the use of autonomous vehicles (AVs). Therefore, this survey discusses the current development of two main components of an ITS: (1) gathering of AVs surrounding data using sensors; and (2) enabling vehicular communication technologies. First, the paper discusses various sensors and their role in AVs. Then, various communication technologies for AVs to facilitate vehicle to everything (V2X) communication are discussed. Based on the transmission range, these technologies are grouped into three main categories: long-range, medium-range and short-range. The short-range group presents the development of Bluetooth, ZigBee and ultra-wide band communication for AVs. The medium-range examines the properties of dedicated short-range communications (DSRC). Finally, the long-range group presents the cellular-vehicle to everything (C-V2X) and 5G-new radio (5G-NR). An important characteristic which differentiates each category and its suitable application is latency. This research presents a comprehensive study of AV technologies and identifies the main advantages, disadvantages, and challenges.

A Survey of Autonomous Vehicles: Enabling Communication Technologies and Challenges.

A predictive safety filter for learning-based control of constrained nonlinear dynamical systems

Recent successes in the field of machine learning, as well as the availability of increased sensing and computational capabilities in modern control systems, have led to a growing interest in learn...

https://www.annualreviews.org/doi/pdf/10.1146/annurev-control-090419-075625

Learning-Based Model Predictive Control: Toward Safe Learning in Control

We consider the problem of predicting the motion of vehicles in the surrounding of an autonomous car, for improved motion planning in lane-based driving scenarios without inter-vehicle communication. First, we address the problem of single-vehicle estimation by designing a filtering scheme based on an Interacting Multiple Model Kalman Filter equipped with novel intention-based models. Second, we augment the proposed scheme with an optimization-based projection that enables the generation of non-colliding predictions. We then extend the approach to the problem of simultaneously estimating multiple vehicles by using a hierarchical approach based on a priority list. The priority list is dynamically adapted in real-time according to a proposed sorting algorithm. Finally, we evaluate the proposed scheme in simulations using real-life vehicle data from the Next Generation Simulation (NGSIM) dataset.

/pdf/interaction-aware-motion-prediction-for-autonomous-driving-a-50pmwitlgl.pdf

Interaction-Aware Motion Prediction for Autonomous Driving: A Multiple Model Kalman Filtering Scheme

This paper presents convex formulations for inverse optimal control problems for linear systems to infer cost function matrices of a quadratic cost from both optimal and non-optimal closed-loop gains. It introduces an optimality measure which enables a formulation of the problem as a convex semidefinite program for the general case and a linear program for several special cases. We derive an explicit algebraic expression for general objective function matrices as well as conditions under which the solution to the inverse optimal control problem is unique. The result is derived by means of a vectorization and parametrization of the algebraic Riccati equation. A simulation example highlights the robust performance in the presence of noise on the measured closed-loop gain and the computational efficiency of the proposed problem formulations.

/pdf/convex-formulations-and-algebraic-solutions-for-linear-28k5r9udn2.pdf

Convex Formulations and Algebraic Solutions for Linear Quadratic Inverse Optimal Control Problems

This brief presents an inverse optimal control methodology and its application to training a predictive model of human motor control from a manipulation task. It introduces a convex formulation for learning both objective function and constraints of an infinite-horizon constrained optimal control problem with nonlinear system dynamics. The inverse approach utilizes Bellman’s principle of optimality to formulate the infinite-horizon optimal control problem as a shortest path problem and the Lagrange multipliers to identify constraints. We highlight the key benefit of using the shortest path formulation, i.e., the possibility of training the predictive model with short and selected trajectory segments. The method is applied to training a predictive model of movements of a human subject from a manipulation task. The study indicates that individual human movements can be predicted with low error using an infinite-horizon optimal control problem with constraints on the shoulder movement.

/pdf/constrained-inverse-optimal-control-with-application-to-a-5orcd69jsv.pdf

Constrained Inverse Optimal Control With Application to a Human Manipulation Task

This paper presents a method for inverse learning of a control objective defined in terms of requirements and their joint probability distribution from data. The probability distribution characterizes tolerated deviations from the deterministic requirements and is learned using maximum likelihood estimation from data. Further, this paper introduces both parametrized requirements for motion planning in autonomous driving applications and methods for the estimation of their parameters from driving data. Both the parametrized requirements and their joint probability distributions are estimated using a posterior distribution such that the control objective is personalized from a prior as driver data are accumulated. Finally, three variants of the learning method are presented that vary in computational complexity and data storage requirements. Key advantages of the proposed inverse learning method are a relatively low computational complexity, a need for a limited amount of data, and that the data do not have to be segmented into specific maneuvers, which makes the method easily implementable. Learning results using data of five human drivers in a simulation environment suggest that the proposed model for human-conscious driving along with the proposed learning method enable a more natural and personalized driving style of autonomous vehicles for their human passengers.

/pdf/inverse-learning-for-data-driven-calibration-of-model-based-y5p9ohfrbi.pdf

Marcel Menner

Papers

Learning-Based Model Predictive Control: Toward Safe Learning in Control

Interaction-Aware Motion Prediction for Autonomous Driving: A Multiple Model Kalman Filtering Scheme

Convex Formulations and Algebraic Solutions for Linear Quadratic Inverse Optimal Control Problems

Constrained Inverse Optimal Control With Application to a Human Manipulation Task

Inverse Learning for Data-Driven Calibration of Model-Based Statistical Path Planning