Control and optimization approaches for power management in energy-aware battery-powered systems

• Main Author: Wang, Tao
• Language: en_US
• Published: Boston University 2015
• Online Access: https://hdl.handle.net/2144/11081

Thesis (Ph.D.)--Boston University === This dissertation is devoted to the power management of energy-aware battery-powered systems (BPSs). Thanks to the popularization of wireless and mobile devices, BPSs are increasingly and widely used. However, the development of BPSs is hindered by the short lifetime of batteries and limited accessibility to charging sources. The first part of this dissertation focuses on the power management of BPSs based on an analytical non-ideal battery model, the Kinetic Battery Model (KBM). How to control discharge and recharge processes of the BPS to optimize the system performance is investigated. Problems for single-battery systems and multi-battery systems are studied. In the single-battery case, the calculus of variations approach gives analytical solutions to the cases with both fully and partially available rechargeability. The results are consistent with the ones derived under a different non-ideal battery model, demonstrating the validity of the solution to the general non-ideal battery systems. In the multi-battery systems, in order to maximize the minimum terminal residual energy among all batteries, the similar methodology is employed to show an optimal policy making equal terminal energy values of all batteries as long as such a policy is feasible, which simplifies the derivations of the solution. Furthermore, the KBM is introduced into a routing problem for lifetime maximization in wireless sensor networks (WSNs). The solution not only preserves the properties of the problem based on an ideal battery model but also shows the applicability of the KBM to large network problems. The second part of the dissertation is focused on BPV systems. First, the energy-aware behavior of electric vehicles (EVs) is studied by addressing two motion control problems of an EV, (a) cruising range maximization and (b) traveling time minimization, based on an EV power consumption model. Approximate controller structures are proposed such that the original optimal control problems are transformed into nonlinear parametric optimization problems, which are much easier to solve. Finally, motivated by the significant role of recharging in BPVs, the vehicle routing problem with energy constraints is investigated. Optimal routes and recharging times at charging stations are sought to minimize the total elapsed time for vehicles to reach the destination. For a single vehicle, a mixed-integer nonlinear programming (MINLP) problem is formulated. A decomposition method is proposed to transform the MINLP problem into two simpler problems respectively for the two types of decision variables. Based on this, a multi-vehicle routing problem is studied using a flow model, where traffic congestion effects are considered are included. Similar approaches to the single vehicle case decompose the coupling of the decision variables, thus making the problem easier to solve.