Introduction

The introduction is still being written and will go here.

Background

Researchers at New Mexico State University and Hashemite University in Jordan reported that 83% of world energy use was supplied by fossil fuels in 2020, while only 5.7% was supplied by the three most prominent renewables: wind, solar, and hydro (Holechek et al.). The same team noted six reasons why renewables are so much harder to integrate than fossil fuels, two of them being intermittency and location mismatch. Our most widely known renewable sources are often not available when and where people need them, and they depend on factors we can’t control. The simplest example is solar: panels can only absorb energy during the day, with direct sunlight. Researchers at Lawrence Berkeley National Laboratory found that utility-scale solar farms only produce 17% to 31% of the power they would make if they ran at full strength every hour of the day (Wiser et al.). Weather patterns and nighttime hours drastically limit how much solar generation is possible. The same lab also found that land-based wind farms reach only 40% of their maximum.

Human motion works differently. It doesn’t depend on weather, daylight, or water levels. It only depends on ordinary people being present. Luckily, we are pretty good at that, especially in airports, gyms, recreational resorts, or any place where people are constantly circulating in large numbers.

To judge what human motion can realistically offer in the future, you need to know some simple electrical measures. A watt describes how fast energy is produced or used, and a kilowatt-hour describes how much energy is used over time: running a 1,000-watt microwave for one hour uses one kilowatt-hour. According to the U.S. Energy Information Administration, the average American home uses 10,791 kilowatt-hours a year, or about 30 kilowatt-hours a day.

The human body generates significantly less than that. Researchers at universities in Poland estimate that the human body generates about 50 to 300 watts while exercising, and only about 20 watts over long exercise periods (Moskwa & Łebkowski). That means an hour of stationary cycling gives at most about 0.3 kilowatt-hours, which doesn’t come close to powering the average home for a day. If the numbers are so drastically different, what makes some people think human motion is plausible as a renewable power source?

The answer starts with how we capture energy. To capture human energy, you need a device that converts motion into electricity. There are multiple ways to accomplish this, but only a few can do it efficiently. Visconti et al., engineers at the University of Salento in Italy, identified the three most widely used methods: piezoelectric, triboelectric, and electromagnetic generation. Piezoelectric materials generate a small electrical charge when they are bent, squeezed, or put under pressure. Triboelectric generation takes advantage of the charge that builds up between two surfaces when they touch or rub against each other quickly (static buildup). Lastly, electromagnetic generation involves spinning a magnet past coils of copper wire, which creates an electric current. Visconti et al. explain that, of the three systems, electromagnetic generation is the strongest. Triboelectricity is hard to justify because it requires a lot of static buildup to generate just a few milliwatts. Piezoelectric materials work most efficiently when vibrating about 200 times per second, but footsteps happen only one to ten times per second. That pace is well within the range where electromagnetic generators can function efficiently. These findings don’t eliminate triboelectric and piezoelectric methods, however. Depending on the application, either may be a better fit for a different environment. For example, electromagnetic induction requires moving parts in order to spin a motor, while piezoelectricity needs no moving parts and is significantly cheaper. As it stands, electromagnetic generation has the highest ceiling for producing wattage, but other factors go into choosing a generation system.

Let’s zoom in on electromagnetic generation that targets floor-mounted systems. Jintanawan et al., engineers at universities in Thailand, tested two electromagnetic floor tile mechanisms: a rack-and-pinion design, which uses toothed bars and gears, and a lead screw design, which uses a threaded rod that spins when the tile is pressed down. Their experiments showed that, when the floor tile was stepped on, the lead screw design produced almost three times as much power within the same compression distance. Why such a big difference? With a lead screw, the generator can keep spinning freely after the tile has been pushed down, until the resistance of the motor eventually makes it stop. With a rack-and-pinion design, the generator is forced to stop moving once the tile has been fully compressed. My own floor tile builds on this lead screw approach, because it can generate significantly more power than the rack-and-pinion method. The energy in human motion is there, and there are ways to capture it. You could look anywhere on the human body and find a way to capture energy from that spot. However, some spots have significantly more potential for generating meaningful power.

How Much Power a Person Can Make

In 1996, Thad Starner, a researcher at the MIT Media Lab, found that a 68-kilogram person walking at a normal two steps per second, with their heel moving about 5 cm above the ground, has roughly 67 watts of power available just from their heel movement. This number differs from what real harvesting devices will generate, because it is a theoretical ceiling. Still, it establishes what humans are capable of. Another researcher from the same Media Lab, Joseph Paradiso, added that the body can produce a few watts from the arms, but tens of watts from the legs. In light of Paradiso’s assertion, lower-body generation, such as pedal- and floor-type generation, should produce significantly more power than upper-body or arm-movement generation.

This data was presented over two decades ago, and it only describes theoretical ceilings. Yet recent experiments at universities have taken those theories and applied them in the lab. Donelan et al., a team from Simon Fraser University in Canada, the University of Michigan, and the University of Pittsburgh, built a 3.5-pound knee brace with an electric generator inside it. The brace works in a similar fashion to regenerative braking in EVs. At the end of each step, your leg swings forward for the next, and your knee needs to slow down before your foot lands, the same way a car needs to use its brakes when coming to a stop. In both the knee and car scenarios, the braking energy normally turns into heat. Instead, the electromagnetic generator inside the brace harvests it, similar to how EVs partially recharge their batteries when they slow down. The researchers reported that the knee-mounted generator produced 4.8 watts at a walking pace of 5.6 kilometers per hour. They concluded that watt-level energy harvesting with electromagnetic generators is practical (Donelan et al.). Their conclusion uses the term “watt-level,” meaning only a few watts. Still, watt-level power is sufficient to charge something like a phone, so the knee brace can supply real, usable power.

A fair question is whether human energy harvesting “steals” your energy. In other words, do these devices force the person to exert more effort? The knee brace results show that they don’t have to. The University of Michigan’s summary of the study said that a typical hand-crank generator takes about 6.4 watts of the user’s own effort to make one watt of electricity, because the muscles and the generator both waste energy (“Knee Brace”). When the brace collected energy only while the knee was braking, the user needed less than one extra watt of effort for every watt produced. The difference is that the brace only collects energy the leg would typically lose as heat, so the user isn’t made to exert more energy than necessary. Rather, it takes advantage of a movement whose energy would otherwise be lost. Kinetic floor tiles apply a very similar concept: they harvest energy from steps people were already planning to take. Still, the resistance added by a floor tile is a possible limit that will be discussed later.

As mentioned earlier, capturing this energy is not limited to electromagnetic induction designs. Zhu et al., a team led by researchers at Georgia Tech, built a triboelectric generator smaller than a human palm. When stepped on, it delivered 1.2 watts for a brief instant, which is enough to power 600 LED bulbs at once. That figure is a very short peak in the middle of a footstep, not a steady average, and a single burst cannot power anything for long. The point is that an entirely different technology can also produce watt-level power from a footstep. If many steps are captured and stored efficiently, the energy from these brief bursts adds up. The team’s results show that technologies beyond electromagnetic designs can still generate meaningful energy.

When you view these studies side by side, it’s clear that human motion is measurable, repeatable, and capturable by more than one method. A few watts is tiny next to a power plant, but every person who crosses a floor adds their own few watts, so the total grows with the crowd. Kinetic energy harvesting becomes competitive when it expands to the scale other renewables operate at. At that point, running sensors, LED lighting, or phone charging at the watt level is easy work. There is a healthy amount of research on kinetic flooring, but that’s only half the picture. What has been applied in the real world to prove this isn’t just a novelty?

The Energy Is Real and Capturable

Let’s look at Pavegen, a British company that makes floor tiles that generate electricity when you step on them. Pavegen reports that its tiles produce about three joules of energy per step, or up to 5 watts while someone is walking. That’s enough to run sensors, LED lights, and screens, or to charge batteries. That said, three joules seems underwhelming. For reference, three joules can light a 1-watt LED for about three seconds. Do you remember when I said this idea becomes competitive when it reaches the scale it needs? Here’s an example of that: the company reports projects in 36 countries with more than 200 installed systems, and says it has captured over 500 million footsteps (“Firm That Turns Footsteps”). These figures do shift between sources, though. For instance, journalist Gareth Huw Davies reported two to four joules per step for the same Pavegen tile. That inconsistency is a reason for caution, but footsteps reliably produce measurable energy, and the output looks significantly more reasonable when tiles are placed all over the world.

Pavegen’s installations range across the world, but I want to point out a few of them. At London’s Heathrow Airport, 51 tiles are placed along a walkway in Terminal 3 and directly power the lights above them (Davies). At Abu Dhabi International Airport, another walkway between terminals has Pavegen tiles covering 16 square meters, which collect energy from about 8,000 people a day (“An Internet of Beings”). Lastly, at the 2013 Paris Marathon, a 25-meter strip of tiles generated 4.7 kilowatt-hours from the runners who crossed it (Devon). These three installations are in areas with heavy, predictable foot traffic. The Paris result also shows that the runners who crossed the strip over the entire marathon produced about one sixth of what the average American home uses in a single day. That is a reason to use this technology for low-power uses: it works in many scenarios, but not as a replacement for the grid.

Researchers at universities have found similar results with their own experimental designs. Salim et al., engineers at universities in Egypt and Saudi Arabia, built a floor tile with 14 piezoelectric sensors and tested it with foot traffic of about 80 people per minute. With a control circuit to steady the voltage, the tile powered two LEDs, and the engineers estimate that 30 tiles covering eight square meters could generate nearly 60 watt-hours over an eight-hour day. This number looks vastly different from Pavegen’s, considering the engineers were working with a piezoelectric tile prototype rather than an electromagnetic system. Their background section explained that piezoelectrics are less suited for footsteps than electromagnetic generators, but they still produced usable power. This test was published in a peer-reviewed journal rather than reported by a company, so it supports the idea behind Pavegen’s claims.

It has already been established that the lower body is the most dominant source to harvest from, but floor tiles are not the only place that lower-body energy can be captured. SportsArt is a fitness equipment company that makes a line of ellipticals, treadmills, and bikes called ECO-POWR. ECO-POWR uses a microinverter, a small device that converts the machine’s electricity into the same kind of power that comes out of a wall outlet. Essentially, you can plug this machine into a wall and its power will flow straight back into the building’s grid. Penn State News reported that the machines are about 74% efficient, meaning 74% of the power a user produces can be returned to the building. Penn State installed ten ellipticals, four recumbent bikes, and two upright bikes in its fitness center (“Penn State Altoona”). SportsArt claims that a single hour-long workout can generate as much as 220 watt-hours (“Sustainable Fitness Solutions”). Other estimates are a bit more modest, but are still real. Moskwa and Łebkowski calculated that a 30-minute moderate cycling session can generate a realistic 62.8 watt-hours. Their conclusion accounts for fatigue and energy loss.

You don’t need a large company to buy this equipment, either. Researchers at San Jose State University modified a regular stationary bike so that its metal flywheel would spin a small generator, successfully charging a battery bank for phones (Huy et al.). They estimated a similar system could be built for $200, compared to the thousand-dollar systems that UC Berkeley, Cal Poly, and the University of British Columbia have added to their gyms (Huy et al.). Even a nightclub has tried this sort of thing. Club Watt in Rotterdam installed a dance floor that turns dancers’ movement into electricity to light the floor itself (“World’s First Ecological Disco”).

It’s apparent that working systems exist in airports, gyms, college campuses, and even a nightclub. None of them claim to power a whole building, but each turns ordinary motion into electricity that serves a real purpose, such as lighting a corridor or feeding power back to a campus grid. For this paper’s goal, that means human kinetic energy is not just a lab experiment. The harder question is how such a small, unsteady power source can be stored efficiently and delivered quickly when it’s needed. That goes a bit deeper into the electronics within these systems.

Making Small Power Usable

Capturing energy is only half of the problem, because the electricity still isn’t ready to use. Salim et al. found that the raw voltage of their piezoelectric floor tiles swung between 0 and 20 volts with each step. A big swing like this would damage most electronics, so they added a circuit that held the output steady at about 5 volts. Lin et al., researchers in China, pointed out a second problem: most motion harvesters have no storage capability. Any energy not used the instant it is generated is lost. A footstep can only produce a short, messy burst of power, and the electronics’ job is to clean up that burst and maintain a steady output efficiently.

There is a typical chain of parts that engineers use to handle this issue. Sarker et al., researchers at the National University of Malaysia and Nebrija University of Spain, describe the process in their review of electromagnetic harvesting circuits. First, a rectifier converts the generator’s output into direct current (DC), electricity that flows in only one direction, which is the kind of power batteries require. Second, a capacitor smooths out the jumps in current from each footstep, and a converter either raises or lowers the voltage to maintain a steady transfer of power. Lastly, you need a charging circuit that feeds the power into a battery or another means of storage. Each step wastes a little energy, depending on the materials, converters, and rectifier methods you use. Choosing the most efficient electrical path is often difficult and costly, but you reap the benefits of higher efficiency. Jintanawan et al. measured their complete rectifier, converter, and battery setup at around 75% efficiency, meaning three quarters of the electricity from the generator made it to storage. Essentially, the electrical system makes all the difference. If your tile can put out 15 watts but the electrical system only passes 50% of that power, you’re losing a significant amount of power you would otherwise be able to generate.

My own floor tile follows the same methodology. The electromagnetic motor inside my prototype works as a three-phase generator. As it spins, it sends electricity out of three contact pads that are soldered to wires. The current in each wire is alternating (AC), meaning it flows in both directions. As mentioned earlier, this power needs to be converted into DC, so I built a circuit board with Schottky diodes. A simple definition of a diode is that it’s a one-way valve for electricity. The diodes let the alternating current through, but force it to flow in only one direction. Think of two alternating-current wires attached to one diode: only one wire comes out the other side, because both alternating currents have been forced into one direct current. When fully put together, this is called a three-phase bridge rectifier. I would also like to point out an issue I came across while building the rectifier. The thin copper traces printed on the circuit board couldn’t carry the full amount of power the floor tile was generating. In other words, only a fraction of the current was able to flow through the diodes. To fix this, I needed another conductive path for the energy to travel. I ended up soldering multiple heavy jumper wires along the underside of the circuit board, so the current had a significantly bigger path to the diodes. The rectifier now works as a proof of concept, and the next steps are the same ones the sources describe in their experiments: a converter to hold the voltage steady, and a charging circuit to connect to a storage method.

Your choice of storage method matters. Riaz et al., researchers at universities in Malaysia, Pakistan, and Spain, compared batteries with supercapacitors. Supercapacitors are small devices that store energy in an electric field instead of through a chemical reaction. Batteries can hold more energy, but they only last around 500 to 1,000 charge cycles. This is a very common issue with lithium-ion batteries today: their capacity depreciates over time, which is why phone and computer batteries lose their life over several years. Supercapacitors, on the other hand, can survive more than 500,000 cycles and recharge in seconds, which is better suited for bursts of power such as a footstep. However, they hold less energy than batteries. For unsteady sources of power like footsteps, the authors noted that a hybrid setup of supercapacitors and batteries is likely the most efficient. This lets the system catch each burst with a supercapacitor, then store it in a traditional lithium-ion battery for the long term. Charging that battery efficiently is its own challenge, and engineers at Texas A&M University have shown it’s possible. Amanor-Boadu et al. built a lithium-ion charging system whose voltage converter reached 90% efficiency and could switch between the harvesting source and a wall outlet. They applied this converter to a heat-based energy harvester, which suggests that a similar converter approach could work for kinetic footstep harvesters.