1. Introduction
Cars and motorcycles are the predominant transportation method in urban environments. This dependence has increased city congestion and levels of air and noise pollution [
1], problems the smart cities/IoT (Internet of things) paradigm seeks to address [
2]. This works provides a solution, integrating bicycling into the smart city platform for monitoring and reducing the pollution from motor vehicles, and as an empowering platform for on-body e-textile autonomous sensors.
Figure 1 shows an integrated smart bicycle concept proposed in this paper for realising this transduction using a switched reluctance motor (SRM) [
3], a harvester generator (SRG) for assisted cycling, and energy harvesting for e-textile sensors. The proposed SRG is cheaper and lightweight when compared to traditional dynamos. It is more environmentally friendly, requiring no permanent magnet, whilst providing greater cyclist comfort, a higher efficiency, and power density. A wireless power transfer (WPT) module enables high-efficiency transmission of the harvested cycling energy to textile-based sensors. A gas sensor, integrated on an e-textile flexible circuit filament, is proposed as a demonstrator.
2. Energy Harvesting from Cycling
Traditional bicycle energy harvesting systems use dynamos. Although dynamos offer good efficiency, they reduce cycling comfort, as the cyclist must contribute significantly more energy to move, which is particularly difficult when cycling uphill. The cyclist’s effort can be reduced by disengaging the dynamo manually prior to strenuous cycling, which is cumbersome. One commercial example of smart bicycle technology is the Copenhagen Wheel [
4]. The system is controlled only through pedaling with one remote for choosing one of three modes, off, motor assist, and pedal assist (regeneration or exercise mode). The wheel harvests energy dissipated during braking and pedaling.
To minimise these disadvantages, a mid-motor type switched reluctance (SR) machine is proposed. The SR machine can be used both as a switched reluctance motor (SRM), and a switched reluctance generator (SRG) that offers assistance to the cyclist for strenuous cycling and harvests energy, respectively. Although the configuration of a magnet and coil is simple and more efficient for energy generation, it cannot function as a motor whilst configured as a generator. This limitation is avoided in the SR machine, enabling it to function as a motor or a generator with good efficiency. The SR machine is tunable, providing a number of choices to the cyclist, for example, turning on the energy harvester when cycling downhill. Harvesting energy in this way would have minimal impact on the cyclist’s comfort, especially with the use of motor assistance during uphill cycling.
A three-phase 18/12 SR machine configuration is considered and modelled in MagNet software to compare the performance of the SR machine on a bicycle based on the cycling profile presented in [
3]. The cyclist power is calculated using
, where
Tav is the average cycling torque over the complete pedal crank revolution and
is the angular speed.
Figure 2 shows the harvested power when the cycling schedule time is arbitrarily fixed at 10% duty cycle, that is, only 10% of cycling power is harvested by the SRG. Results indicate that a cyclist harvests more energy during uphill cycling because the cyclist pedals more and expends more energy in this phase.
3. Wearable Wireless Power Transfer
Flexible textile coils have been previously demonstrated for on- and in-body WPT with high efficiency over a short distance, for wearable and body-planar applications [
5,
6]. To transfer the power wirelessly to the user, near-field non-radiative WPT is proposed using inductive coupling [
6,
7]. Conductive textile threads (40 µm thick) were embroidered into a woven-polyester fabric substrate (280 µm thick), forming planar spiral coils for transmitting and receiving power.
The inductance of the coil is measured using an impedance analyzer and tuned to resonate at 6.78 MHz (the license-free band used for most WPT applications) based on
. Two identical coils are proposed for the bicycle’s handle and for the cyclist’s glove in
Figure 3. The measured coil inductance and resistance at 1 MHz are 4.9 µH and 0.480 Ω, respectively.
A Vector Network Analyzer (VNA) was used to measure the WPT efficiency in
Figure 4 at 6.78 MHz as a function of coil-separation and frequency. The coil demonstrates the highest power transfer efficiency at the selected frequency, due to the lower losses at the low-MHz frequency bands. The reduced WPT efficiency at 0.1 cm separation is a result of detuning due to the coils’ mutual inductance affecting the series inductance, requiring a frequency-tracking circuit at the transmitter. However, more than 80% WPT efficiency is still maintained. Although ultrathin threads of relatively high series resistance were used, the achieved WPT efficiency is similar to state-of-the-art wearable WPT systems [
5,
6]. As the proposed coils are optimized for short-range WPT, high efficiency can be achieved on-body due to the minimised magnetic field dissipation through the human body. Moreover, higher power levels can be transferred compared to radiative power transfer, due to the lower human absorption, compared to microwave frequencies, and the higher regulation’s transmitted power limits.
Table 1 shows a comparison of the proposed WPT coils with reported wearable WPT systems.
4. Requirements for Wearable Sensing
Wearable and textile on-body sensing have attracted recent interest [
8]. A textile-based super capacitor has been presented for storing harvested energy for wearable sensors [
9]. In addition, flexible Radio Frequency (RF) front ends for wireless nodes have been specifically designed for on-off-body communication using standard license free protocols such as Bluetooth Low Energy [
10]. In this work, we investigated a novel wearable gas sensor (
Figure 5), fabricated based on the fabrication process in [
8]. The sensor was implemented for mobile air pollution sensing on-body using commercial gas sensors packaged for weaving in textiles. The sensor runs at 88 mW. At an 80% WPT efficiency, the average power from the SRG from
Figure 2 is 824 mW; sufficient to power multiple gas sensors or other low-power wearable sensors.
5. Conclusions
In this paper, a complete smart-cycling system architecture is proposed. The smart cycling enables the connectivity among cyclists in smart cities. A self-powered sensing system that can be integrated easily to any type of bicycle is considered. To achieve this, a switched reluctance generator to harvest the surplus cycling energy, which is then reused to power on-body sensors, is presented. Moreover, a textile-based wireless power transfer mechanism using inductive coupling and pollution sensing using commercial gas sensors are presented to enable additional use-cases of the harvested cycling energy, empowering crowdsourcing air pollution data through cyclists in smart cities. Future studies will advance from the modular implementation of the architecture proposed in this paper to a full system integration of the SRM/SRG, the inductive power transfer (IPT) and the wearable sensor.