Stochastic MPC With Learning for Driver-Predictive Vehicle Control and its Application to HEV Energy Management

Cairano, Stefano Di; Bernardini, Daniele; Bemporad, Alberto; Kolmanovsky, Ilya

doi:10.1109/tcst.2013.2272179

Cited by 384 publications

(165 citation statements)

References 32 publications

Supporting

Mentioning

165

Contrasting

Order By: Relevance

“…Notice that the values reported for u res represent the percentage ([0,1]) of time during the 20 minutes sampling interval in which the resistor is on, as resulting from the implemented duty cycle. In general, the pattern of the control variables is difficult to be interpreted a posteriori, since it depends in a non-intuitive way on the solution of (20). For this specific case, however, the evolution of u grid can be easily interpreted: due to a relatively mild weather, the battery can easily power the load while being charged, which allows it to provide energy at night as well.…”

Section: Resultsmentioning

confidence: 99%

“…In case of colder days, the controller would instead impose frequent changes in the value of u grid , as shown in the simulations in [26]. The distribution of the computation times needed to solve (20) is shown in the histogram in Fig. 7.…”

Section: Resultsmentioning

confidence: 99%

“…On the other hand, MPC is also increasingly used for the optimal schedule of energy sources in different application domains, such as hybrid electric vehicles [19,20]. The use of MPC for overall energy management of buildings, where the thermal management is considered as one of many factors, has been analyzed, for instance, in [21].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Optimal energy management of a small-size building via hybrid model predictive control

Khakimova

Kusatayeva

Shamshimova

et al. 2017

Energy and Buildings

Self Cite

View full text Add to dashboard Cite

This paper presents the design of a Model Predictive Control (MPC) scheme to optimally manage the thermal and electrical subsystems of a small-size building ("smart house"), with the objective of minimizing the expense for buying energy from the grid, while keeping the room temperature within given time-varying bounds. The system, for which an experimental prototype has been built, includes PV panels, solar collectors, a battery pack, an electrical heater in a thermal storage tank, and two pumps on the solar collector and radiator hydraulic circuits. The presence of binary control inputs together with continuous ones naturally leads to using a hybrid dynamical model, and the MPC controller solves a mixed-integer linear program at each sampling instant, relying on weather forecast data for ambient temperature and solar irradiance. The procedure for controller design is reported with focus on the specific application, and the proposed method is successfully tested on the experimental site.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Optimal energy management of a small-size building via hybrid model predictive control

Khakimova

Kusatayeva

Shamshimova

et al. 2017

Energy and Buildings

Self Cite

View full text Add to dashboard Cite

show abstract

“…Because the Dyna-based control policy is extremely close to the DP-based optimal control policy, the Dyna algorithm has the potential to realize a real-time control strategy in the future. When the present power request is considered a continuous system, the next power request of a vehicle can be predicted accurately using the method introduced in [31,32]. Subsequently, when the power request is combined with the Dyna algorithm, the reward function and transition probability matrix can be updated.…”

Section: Comparative Analysis Of the Results Of Dyna Algorithm Sdp mentioning

confidence: 99%

Reinforcement Learning–Based Energy Management Strategy for a Hybrid Electric Tracked Vehicle

Zou

Liu

et al. 2015

Energies

View full text Add to dashboard Cite

This paper presents a reinforcement learning (RL)-based energy management strategy for a hybrid electric tracked vehicle. A control-oriented model of the powertrain and vehicle dynamics is first established. According to the sample information of the experimental driving schedule, statistical characteristics at various velocities are determined by extracting the transition probability matrix of the power request. Two RL-based algorithms, namely Q-learning and Dyna algorithms, are applied to generate optimal control solutions. The two algorithms are simulated on the same driving schedule, and the simulation results are compared to clarify the merits and demerits of these algorithms. Although the Q-learning algorithm is faster (3 h) than the Dyna algorithm (7 h), its fuel consumption is 1.7% higher than that of the Dyna algorithm. Furthermore, the Dyna algorithm registers approximately the same fuel consumption as the dynamic programming-based global optimal solution. The computational cost of the Dyna algorithm is substantially lower than that of the stochastic dynamic programming.

show abstract

“…The MPC strategies are interesting as they naturally take prediction information into account. The prediction information can be stochastic as in [21,22] or deterministic as in [18][19][20][23][24][25]. The future power demand for the deterministic MPC strategies can be predicted as a function of the current power demand, e.g., the power demand is exponentially decaying in [23,24] and held constant in [25].…”

Section: Introductionmentioning

confidence: 99%

Real-Time Distributed Economic Model Predictive Control for Complete Vehicle Energy Management

et al. 2017

View full text Add to dashboard Cite

Abstract:In this paper, a real-time distributed economic model predictive control approach for complete vehicle energy management (CVEM) is presented using a receding control horizon in combination with a dual decomposition. The dual decomposition allows the CVEM optimization problem to be solved by solving several smaller optimization problems. The receding horizon control problem is formulated with variable sample intervals, allowing for large prediction horizons with only a limited number of decision variables and constraints in the optimization problem. Furthermore, a novel on/off control concept for the control of the refrigerated semi-trailer, the air supply system and the climate control system is introduced. Simulation results on a low-fidelity vehicle model show that close to optimal fuel reduction performance can be achieved. The fuel reduction for the on/off controlled subsystems strongly depends on the number of switches allowed. By allowing up to 15-times more switches, a fuel reduction of 1.3% can be achieved. The approach is also validated on a high-fidelity vehicle model, for which the road slope is predicted by an e-horizon sensor, leading to a prediction of the propulsion power and engine speed. The prediction algorithm is demonstrated with measured ADASIS information on a public road around Eindhoven, which shows that accurate prediction of the propulsion power and engine speed is feasible when the vehicle follows the most probable path. A fuel reduction of up to 0.63% is achieved for the high-fidelity vehicle model.

show abstract

Stochastic MPC With Learning for Driver-Predictive Vehicle Control and its Application to HEV Energy Management

Cited by 384 publications

References 32 publications

Optimal energy management of a small-size building via hybrid model predictive control

Optimal energy management of a small-size building via hybrid model predictive control

Reinforcement Learning–Based Energy Management Strategy for a Hybrid Electric Tracked Vehicle

Real-Time Distributed Economic Model Predictive Control for Complete Vehicle Energy Management

Contact Info

Product

Resources

About