Closed‐loop hierarchical control strategies for connected and autonomous hybrid electric vehicles with random errors

Hierarchical control strategies for energy management of connected hybrid electric vehicles in urban roads

<div class="section abstract"><div class="htmlview paragraph">This paper presents a hierarchical control architecture to coordinate a group of connected vehicles on signalized intersection roads, where vehicles are allowed to change lane to follow a prescribed path. The proposed hierarchical control strategy consists of two control levels: a high level controller at the intersection and a decentralized low level controller in each car. In the hierarchical control architecture, the centralized intersection controller estimates the target velocity for each approaching connected vehicle to avoid red light stop based on the signal phase and timing (SPAT) information. Each connected vehicle as a decentralized controller utilizes model predictive control (MPC) to track the target velocity in a fuel efficient manner. The main objective in this paper is to consider mandatory lane changes. As in the realistic scenarios, vehicles are not required to drive in single lane. More specifically, they more likely change their lanes prior to signals. Hence, the vehicle decentralized controllers must prepare to cooperate with the vehicle that has a mandatory lane change request (host vehicle). The cooperative mandatory lane change is accomplished by inserting a virtual vehicle on the host vehicle’s target lane. The simulation results show the advantage of our proposed approach on both the lane change duration and vehicle fuel economy.</div></div>

This paper presents the open-source stochastic model predictive control framework GRAMPC-S for nonlinear uncertain systems with chance constraints. It provides several uncertainty propagation methods to predict stochastic moments of the system state and can consider unknown parts of the system dynamics using Gaussian process regression. These methods are used to reformulate a stochastic MPC formulation as a deterministic one that is solved with GRAMPC. The performance of the presented framework is evaluated using examples from a wide range of technical areas. The experimental evaluation shows that GRAMPC-S can be used in practice for the control of nonlinear uncertain systems with sampling times in the millisecond range.

ABSTRACTThis article proposes a parallel algorithm of stochastic model predictive control (SMPC) for linear discrete‐time systems. The proposed algorithm uses multiple threads to solve respective optimization problems at different time steps in parallel, without inter‐thread communication during each optimization. Each thread sequentially solves a stochastic optimal control problem (SOCP) for optimizing closed‐loop performance against future disturbances and a partial minimization problem for optimizing a subset of decision variables of the SOCP. The partial minimization problem computes the control input according to the last measured state to satisfy chance constraints and can be solved much faster than the original SOCP. Consequently, the proposed algorithm can reduce the implementable minimum sampling period compared to the original SMPC. A numerical example of controlling room temperature showed that the proposed algorithm achieves control performance almost identical to that of closed‐loop SMPC, which optimizes closed‐loop performance. However, the implementable minimum sampling period of the proposed algorithm is much shorter than that of the closed‐loop SMPC, even with parallelization overhead.

A methodology has been developed for optimizing building supervisory control strategies, employing building models that incorporate stochastic models of occupant behaviour and serve as the objective function evaluator in a stochastic model predictive control (SMPC) architecture. The SMPC architecture accounts for variability in building performance due to occupant behaviour and is shown to generate a sequence of automatic window opening decisions for a mixed mode building which lead to more robust building performance in the face of occupant window use than a heuristic controller. A set of receding optimization time horizons are described which enable the use of complex building models in simulated SMPC. Results of a case study show that deterministic optimization predicts a 50% increase in building performance, while stochastic optimization leads to a more conservative and more reliable 33% performance improvement, which takes into consideration the impact of occupant behaviour.

Chapter 14 - Fast stochastic model predictive control of end-to-end continuous pharmaceutical manufacturing

A stochastic model predictive control (SMPC) method is proposed for trajectory tracking of quadrotor helicopters with stochastic disturbances. The stochastic disturbances are assumed to be Gaussian white noises. The existing robust control methods generally guarantee that the tracking performances under the worst-case disturbances satisfying given conditions. However, the worst-case disturbances may have a vanishingly small probability of occurrence in practice, so these methods show great conservative. In this paper, the tracking errors caused by the disturbances are limited in the acceptable range with the given probability. The hierarchical control structure is used. In the outer loop, the position loop, SMPC controllers are designed for position tracking. In the inner loop, the attitude loop, the nonlinear dynamics are linearized by the feedback linearization strategy, and an SMPC controller is designed for attitude tracking. Finally, simulation results verify the effectiveness of the proposed method.

Stochastic Model Predictive Control (SMPC) has attracted increasing attention for autonomous driving in recent years, since it enables collision-free maneuvers and trajectory planning and can deal with uncertainties in a non-conservative way. Many promising strategies have been proposed on how to use SMPC to select appropriate maneuvers and plan safe trajectories in uncertain environments. The limitation of these approaches is that they focus on scenarios where only one vehicle is controlled by SMPC and is, thus, reacting to the surrounding vehicles; however, the surrounding vehicles do not react to the SMPC-controlled vehicle, which means there is no mutual interaction. However, when multiple autonomous vehicles are driving on the road, each individual vehicle will take the behavior of the other surrounding vehicles into account and adjust its individual decisions accordingly in trajectory planning. This paper, therefore, examines in simulations how the interactive control system of multiple SMPC-controlled vehicles behave based on a Distributed SMPC (DSMPC) framework. For a three-lane highway scenario, we first investigate the effects of the risk parameter of the collision avoidance probabilistic constraint on non-interactive and interactive vehicle systems and provide insights into how to parameterize the controllers in interactive vehicle systems.

This article describes the design, implementation, and evaluation of an active lane change control algorithm for autonomous vehicles with human factor considerations. Lane changes need to be performed considering both driver acceptance and safety with surrounding vehicles. Therefore, autonomous driving systems need to be designed based on an analysis of human driving behavior. In this article, manual driving characteristics are investigated using real-world driving test data. In lane change situations, interactions with surrounding vehicles were mainly investigated. And safety indices were developed with kinematic analysis. A safety indices–based lane change decision and control algorithm has been developed. In order to improve safety, stochastic predictions of both the ego vehicle and surrounding vehicles have been conducted with consideration of sensor noise and model uncertainties. The desired driving mode is decided to cope with all lane changes on highway. To obtain desired reference and constraints, motion planning for lane changes has been designed taking stochastic prediction-based safety indices into account. A stochastic model predictive control with constraints has been adopted to determine vehicle control inputs: the steering angle and the longitudinal acceleration. The proposed active lane change algorithm has been successfully implemented on an autonomous vehicle and evaluated via real-world driving tests. Safe and comfortable lane changes in high-speed driving on highways have been demonstrated using our autonomous test vehicle.

The initial problem on building a mobile robot using embedded system is the hardware flexibility. Currently most of mobile robots use microcontroller. The weakness of building mobile robot using microcontroller is the form and the specification which is inflexible to be added or changed. For that reason this research use FPGA soft processor to design a hardware architecture in the form of system on chip (SoC) as the motor velocity control and robot heading. We divide the mobile robot control into 2 levels which are high level and low level control. High level control is including the positioning and path tracking, while low level control is including kinematic, speed control, heading lock and communication. This research focused on low level control design. The achieved result with the use of FPGA soft processor, is that the programmer can flexibly add and change the hardware architecture and faster program execution resulted from parallel process. On the motor velocity testing, the time to get back into the steady state after receiving disturbance was about 0.6 seconds. On the heading lock testing, the time of the robot to get back into the set point is around 1 second. Resources required to implement soft processor system are 3883 slices register, 1130 slices flip-flop, 3825 slices LUTs and 191 RAM Blocks.

Due to the features of strong non-linearity and hybrid driving with multi-power sources, a novel deep reinforcement learning(DRL)-based hierarchical energy control architecture is presented for a group of connected hybrid electric vehicles (HEVs) in the platoon through cloud platform. Firstly, a higher-level model predictive control (MPC) law is designed to determine the desired acceleration of connected HEVs in a platoon. Secondly,a reward function with the change of state of charge (SOC) and instantaneous fuel consumption as independent variables is constructed, and the expert knowledge-based deep deterministic policy gradient (DDPG) method with prioritized experience replay (PER) is used for designing the lower-level energy management strategy of connected HEVs platooning. Then, to illustrate the rationality of decision-making by the presented DDPG strategy, the influence of state variables such as vehicle speed, desired acceleration and battery SOC on the agent action values of presented DDPG method is discussed. Finally, the test results show that the proposed control scheme can reasonably allocate the engine and motor powers in real time, and finally achieve the safe and energy-saving driving of connected HEVs in the platoon.

In this paper a hierarchical model predictive control framework for the energy management of a hybrid electric vehicle is extended by replacing the supervisory linear model predictive controller with a nonlinear model predictive controller. The nonlinear system dynamics of a parallel hybrid electric vehicle are incorporated into a nonlinear optimal control problem which is solved at each time instance. The discretization of the continuous time system dynamics is realized via direct multiple shooting. The low-level linear model predictive controller considers the actuator dynamics for optimal tracking of the driver's torque demand while the high-level controller considers the slow dynamics of the vehicle and the battery. The proposed strategy is validated in simulation with the worldwide harmonized light vehicles test cycle and compared to a hierarchical model predictive control framework with linearized system dynamics.

In this paper a novel method is presented for hierarchical control of autonomous vehicles consisting of high-level (HLC) and low-level (LLC) control systems. The LLC system can be commanded to assure a prescribed velocity and steering profile using PID-type active suspension, velocity and steering control. The outputs of the LLC system are the generalized forces and moments required by the vehicle. Because of its nature, the LLC system cannot eliminate position and orientation errors caused by the unknown centre of gravity point, drifting and time-varying friction. These errors are compensated by the HLC which uses the kinematic model of the vehicle to generate the reference trajectory from the prescribed profiles and employs a linear quadratic optimal controller to compensate the tracking errors. The control algorithm minimizes the perturbations and the optimal solution is realized as a time-varying state-feedback control law. The efficiency of the hierarchical vehicle control system is validated for realistic motion maneuvers by simulation on the nonlinear vehicle model using multibody robotic formalism.

Due to the uncertainty of human driving behavior, an autonomous vehicle (AV) may activate its emergency protection policy frequently and unnecessarily when interacting with human drivers, degrading driving efficiency. To address this issue, this paper proposes an optimization framework that coordinates nominal motion planning and emergency protection control, using stochastic model predictive control and a sparse scenario tree to anticipate future collision events. By explicitly considering future safety risks, AV's planner can reduce unnecessary emergency interventions and achieve a balance between safety and performance. The proposed framework integrates backward reachability analysis to precompute the set of unsafe states, and employs recursive Bayesian estimation with a Boltzmann noise-rational human model to predict driver behavior. Simulations demonstrate that the proposed control reduces control costs by 13.1% and improves efficiency by 63.4% compared to baseline methods. Moreover, the interactive experiment platform based on two minicars demonstrates that the proposed method is effective in real-time environments.

For heavy-duty diesel engines, NOX emissions reduction is strongly constrained by fuel efficiency. This paper presents a hierarchical model predictive controller (H-MPC) for coordinated control of tailpipe NOX emissions and fuel consumption. The H-MPC uses the separation of slow and fast dynamics that exist in the engine and its aftertreatment system. The controller is synthesized with an architecture in which a high-level MPC uses a longer prediction horizon compared to the low-level predictive controller which tracks the high-level controller command and manages the thermal dynamics of the aftertreatment system. Engine load preview enables the high-level controller to estimate the desired catalyst temperature ahead of time and addresses the selective catalytic reduction (SCR) slow thermal dynamics. Calculated by the high-level controller, the intake manifold pressure, and the start of injection (SOI) crank angle is used as reference trajectories in the low-level controller that regulates fast dynamical behaviors such as engine out NOX emissions. Hardware-in-the-loop (HIL) validation of this integrated H-MPC on a rapid prototype controller shows that when the SCR catalyst temperature is above light-off temperature (warmed-up condition), the engine operation is shifted to operate with the best fuel economy since the warmed-up SCR can efficiently reduce the engine-out NOX emissions. Results indicate that up to 0.8% benefit in cycle averaged BSFC along with a 13% reduction in tailpipe NOX compared to a stock engine calibration can be achieved with the coordinated engine and aftertreatment system through H-MPC.