The power sources must be considered as well. In particular, the energy delivered by a battery depends on the rate at which it is consumed. Consequently, reducing peak power can increase the battery life of the system by increasing the energy available. Large peak power has been shown to reduce the energy delivered by a battery by up to 40 percent. Electrochemical capacitors can be used to mitigate peak demands and improve battery life by up to 10 percent. The use of energy-aware operating system schedulers can reclaim even more of the battery capacity (Martin et al., 2003).

Table 6-1 summarizes mitigation techniques in key areas to improve energy efficiency. It lists improvements that could be realized by using a system approach toward mitigating energy issues associated with just the communications and computation functions of the Land Warrior.

In summary, greedy approaches to energy efficiency that consider subsystems in isolation will not be optimal. A system-level approach, one that considers energy consumers and power sources, is the proper method for examining energy efficiency in battery-powered computing systems.

POWER MANAGEMENT AND DISTRIBUTION

Land Warrior systems can be subdivided into four functional areas: displays, computing, sensors, and communications. Each functional area requires one or more power sources for electronics such as were listed in Table 5-1.

Distributed vs. Centralized

There are two basic approaches to power distribution: centralized and distributed. In centralized power distribution there is a central power source that is distributed by wires to the various power sinks. The centralized source may generate bulk power to distributed power regulators, which smooth out spikes and maintain voltage at the specified level for the various sinks (or for small, local, rechargeable batteries—this would allow various elements of the system to operate briefly if the centralized power source went down). If the system is small enough and the sinks have similar voltage requirements, the regulation may also be done at the centralized source. Since the energy density of batteries usually increases with battery size (owing to less packaging per unit volume), the centralized power option should result in higher energy density than the distributed power option and a lighter weight battery for the soldier.

The centralized power option should allow one to reduce the types of power sources needed. In theory, by coupling with DC-DC converters, it should be possible to use only one type of power source, with a secondary one of similar type as backup. Owing to the inefficiencies of DC-DC converters, there is a penalty of about 10 percent, but energy can be saved in the sink by keeping the voltage constant at the lowest possible level, particularly where the battery voltage changes greatly during discharge.

Additionally, central power will require that equipment be tethered to the central power source. The tethered equipment that would probably cause the most practical problems for the soldier is the weapon subsystem and the helmetmounted heads-up display (HUD). The proposed OFW weapon subsystem has 14 components that draw power (Acharya, 2003). However, because a tethered weapon is prone to becoming entangled with protruding objects, its reliability should be studied in the field before implementation. Ultimately, the need to exchange an enormous amount of data via wires will dictate that the four functional areas of the OFW system be tethered together through wires. Consequently, one should be able to distribute the power from a central source through these tethered wires also. In other words, the unavailability of a wireless body LAN necessitates the use of tethered wires, which, in turn, dictates the centralized power approach. Advantages and disadvantages are summarized in Table 6-2.

In distributed power generation, the generation source, power regulator, and sinks are distributed around the body. There is no need for encumbering, tethering wires to connect

TABLE 6-1 Techniques for Mitigating Energy Issues in Key Land Warrior System Components and Improvements That Could Be Realized

Component

Mitigation Technique

Improvement

Power source

Battery

Reduce peak draw

Up to 10% more available energy

Power sink

Communications

Energy-aware network routing

Local processing

Up to 50% fewer hops, 50% less energy

Delineation of local versus remote processing based on communications/processing cost

Computation

Remote processing

Dynamic CPU speed setting

Delineation of local versus remote processing based on communications/processing cost

Prediction of idle time and active power within 5% of actual

NOTE: CPU, central processing unit.

SOURCE: Adapted from Martin et al., 2003.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement