Parasitic Power Harvesting in Shoes

system to date has served all of the needs of wearable computing—light weight, minimum effort, high power generation, convenient power delivery, and good power regulation. We believe that our approach has the potential to solve these problems for a class of wearable devices by placing both the generator and powered electronics in a location where considerable energy is easily available, namely the shoe. As the power requirements for microelectronics continue decreasing, environmental energy sources can begin to replace batteries in certain wearable subsystems. In this spirit, this paper examines three different devices that can be built into a shoe, (where excess energy is readily harvested) and used for generating electrical power "parasitically" while walking. Two of these are piezoelectric in nature: a unimorph strip made from piezoceramic composite material and a stave made from a multilayer laminate of PVDF foil. The third is a shoe-mounted rotary magnetic generator. Test results are given for these systems, their relative merits and compromises are discussed, and suggestions are proposed for improvements and potential applications in wearable systems. As a self-powered application example, a system had been built around the piezoelectric shoes that periodically broadcasts a digital RFID as the bearer walks. In previous studies [5], it has been calculated that up to 67 Watts of power are available from heel strikes during a brisk walk (68 kg person, 2 steps/sec, heel moving 5 cm). This level of power extraction from walking would certainly interfere greatly with one's gait. Our philosophy, in contrast, has been to try to generate power entirely parasitically, that is through mechanisms that capture and make use of energy normally dissipated wastefully into the environment. There is much less energy of this type than available through deliberate means of harvesting human power (e.g. through a hand crank or foot pedal), but it is our goal to unobtrusively collect energy for low-power applications. We have approached this problem by using the energy from the weight transfer during a step to perform useful work. 1: Introduction As wearable electronic devices evolve and proliferate, there will be a growing need for more power delivery to distributed points around the human body. Today, much of that storage is provided by batteries and power delivery is via wires. The current approach to power distribution is clearly becoming problematic -as more appliances are carried, we are forced to either use more small batteries that require replacement everywhere or run wires through our clothing to supply appliances from a central power source. Both are undesirable. A better solution is clearly to generate power where it is being used, bypassing the storage and distribution problem altogether. As power requirements drop for most wearable devices, it is no longer infeasible to harvest a useful amount of energy "parasitically" from a normal range of human activity. 2: Background Information The context in which we place our generator is that of a sport sneaker. This type of shoe differs from ordinary shoes in one important feature—its energy dissipating sole. While walking in ordinary "hard" shoes, the foot is rapidly decelerated from its relatively high downward speed to zero velocity relative to the ground—an action that requires the application of relatively large and sudden forces to the foot. Barring shock absorption in the feet, this can be simply modeled as a sudden step in velocity; the force applied to the foot to achieve this deceleration is an impulse (Fig. 1a). Many attempts have been made in the past to tap this source, leading to the consideration of a host of technologies [1] ranging from the construction of various electromechanical generators [2,3] to the surgical placement of piezoelectric material in animals [4]. No generating This impulse causes the foot to decelerate suddenly while the rest of the body is still moving. The force that stops the rest of the body’s mass is transmitted through the legs and compresses the knees and other joints. The ____________________________________________________________________________________ Draft 2.0, August, 1998; Presented at the Second IEEE International Conference on Wearable Computing Figure 1: Dynamics for hard vs. cushioned foot strikes function of the insole and midsole in the sport sneaker is to work as a low-pass filter for this step in velocity, reducing the amount of force applied to the joints (Fig. 1b). This reduces any stress that the joints experience and also reduces the incidence of sports injuries. Figure 3: Layout of the PVDF Power Insole 3: System Descriptions One obvious means of parasitically tapping energy in this context is to harness the bending of the sole, which is attempted in our first system. This is a laminate of piezoelectric foil, shaped into an elongated hexagon, as shown in Fig. 3. This “stave” is a bimorph built around a central 2-mm flexible plastic substrate, atop and below which are sandwiched 8-layer stacks of 28-micron PVDF (polyvinylidineflouride) sheets [6], epoxy-bonded as shown. This stave was designed in collaboration with K. Park and M. Toda of the Sensor Products Division of Measurement Specialties (formerly AMP Sensors) [7]; its shape was chosen to conform to the footprint and bending distribution of a standard shoe sole. As the stave is very thin (under 3 mm), it can be easily molded directly into a shoe sole. When the stave is bent, the PVDF sheets on the outside surface are pulled into expansion, while those on the inside surface are pushed into contraction (due to their differing radii of curvature), producing voltages across silver-inked electrodes on each sheet through the dominant "3-1" longitudinal mode of piezoelectric coupling in PVDF. In order to lower the impedance, the electrodes from all foil sheets are connected in parallel (switching polarities between foils on opposite laminate surfaces to avoid cancellation), resulting in a net capacitance of 330 nf. An actual stave used in our tests is shown in Fig. 4. The result is that the force and displacement values over time for the bottom and top of the midsole are not the same—as in any passive filter, there is an energy loss in the sole while it performs this filtering function. The energy lost is in the higher harmonics of the step and is dissipated through internal losses in the sole. When the sole springs back after the step it does not exert as much force as before, returning less energy than was put into it, and it is this energy that we are trying to capture (Fig. 2). The energy obtained from the shoe is not free—as the harvested power grows, there is a noticeable additional load as the shoe demands more energy to be put into it while supplying less restoring force (somewhat like walking on sand). Our systems strive to make this burden beneath notice, ideally loading the user’s stride exactly as much as common sport shoes today. Figure 2: Force/displacement curve for a sneaker sole Another promising mode of harnessing parasitic power in shoes is to exploit the high pressure exerted in a heel strike. There are many ways to tap this; for instance some groups [8] are trying to develop highly elastic