Kai Geissdoerfer
Kai Geissdoerfer has a total of 14 co-authors. Here are the top co-authors by number of shared publications:
- Marco Zimmerling
TU Darmstadt, Darmstadt, Germany
6 shared publications - Ingmar Splitt
TU Darmstadt, Darmstadt, Germany
2 shared publications - Muhammad Moid Sandhu
AEHRC, CSIRO, Brisbane, Queensland, Australia and Queensland University of Technology (QUT), Brisbane, Queensland, Australia
2 shared publications - Raja Jurdak
Queensland University of Technology
2 shared publications - Sara Khalifa
Queensland University of Technology (QUT), Brisbane, Queensland, Australia
2 shared publications - Branislav Kusy
DATA61, CSIRO
1 shared publication
Latest Publications
- Open Access
- 22 May 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Internet of Things
- Vol. 7 (4) , 1-31
KineticWear: Battery-Free On-Device Human Activity Recognition for Wearables
Wearable devices are becoming more prevalent in people’s daily lives, particularly in applications such as activity recognition, health monitoring and fitness management. However, the majority of existing wearable devices remain heavily dependent on battery-based power sources, which introduces several practical and sustainability challenges. Frequent battery replacement or recharging imposes inconvenience on users, increases long-term operational costs, and contributes to environmental concerns associated with battery disposal and resource consumption. To address these issues, we present KineticWear, the first battery-free wearable system that utilises kinetic energy harvested from human activities both as the sole energy source and as a sensing signal for on-device human activity recognition (HAR). Based on a careful end-to-end design of all hardware and software components, KineticWear achieves real-time HAR on an ultra low-power microcontroller unit (MCU) including on-board classification and transmission of the inferred activity over a wireless link. Using empirical data, we find that decision tree (DT) and convolutional neural network (CNN) models offer activity recognition accuracies of 87 % and 99.5 %, respectively. Systematic real-world experiments demonstrate that KineticWear harvests sufficient energy to operate the wearable device up to 95.2 % of the time, and that the device can infer and report an ongoing activity within 8 seconds using DT classification algorithm, taking three orders of magnitude shorter classification time than CNN. Thus, KineticWear offers significantly enhanced performance compared with state-of-the-art off-device activity recognition systems powered by kinetic energy harvesting.
- 22 May 2026
- Published by Association for Computing Machinery (ACM) in ACM Transactions on Internet of Things
- Vol. 7 (4) , 1-31
Open AccessKineticWear: Battery-Free On-Device Human Activity Recognition for Wearables
Wearable devices are becoming more prevalent in people’s daily lives, particularly in applications such as activity recognition, health monitoring and fitness management. However, the majority of existing wearable devices remain heavily dependent on battery-based power sources, which introduces several practical and sustainability challenges. Frequent battery replacement or recharging imposes inconvenience on users, increases long-term operational costs, and contributes to environmental concerns associated with battery disposal and resource consumption. To address these issues, we present KineticWear, the first battery-free wearable system that utilises kinetic energy harvested from human activities both as the sole energy source and as a sensing signal for on-device human activity recognition (HAR). Based on a careful end-to-end design of all hardware and software components, KineticWear achieves real-time HAR on an ultra low-power microcontroller unit (MCU) including on-board classification and transmission of the inferred activity over a wireless link. Using empirical data, we find that decision tree (DT) and convolutional neural network (CNN) models offer activity recognition accuracies of 87 % and 99.5 %, respectively. Systematic real-world experiments demonstrate that KineticWear harvests sufficient energy to operate the wearable device up to 95.2 % of the time, and that the device can infer and report an ongoing activity within 8 seconds using DT classification algorithm, taking three orders of magnitude shorter classification time than CNN. Thus, KineticWear offers significantly enhanced performance compared with state-of-the-art off-device activity recognition systems powered by kinetic energy harvesting.
- Open Access
- 23 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 236-248
Shepherd Nova: A Public Testbed for Rigorous Experiments Under Repeatable Energy-Harvesting Conditions
- 23 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 236-248
Open AccessShepherd Nova: A Public Testbed for Rigorous Experiments Under Repeatable Energy-Harvesting Conditions
- Open Access
- 4 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 883-884
Demo: Battery-free TinyML Made Easy with Riotee
This demo uses a machine-learning-based hot word detection application to showcase the capabilities of Riotee, an open-source and commercially available hardware-software platform for the battery-free Internet of Things. We describe the Riotee hardware consisting of a base module, a debug probe for easy firmware updates, and several expansion boards that enhance functionality without the need for custom-designed printed circuit boards (PCBs). The demo features the classification of live audio recordings using TinyML deep neural network inference aboard a Riotee device. The Riotee device transmits the classification results via Bluetooth Low Energy (BLE) to a smartphone given to visitors. Visitors can also observe how the Riotee software checkpoints and restores critical state in case of power failures via the visualization of logic analyzer traces.
- 4 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 883-884
Open AccessDemo: Battery-free TinyML Made Easy with Riotee
This demo uses a machine-learning-based hot word detection application to showcase the capabilities of Riotee, an open-source and commercially available hardware-software platform for the battery-free Internet of Things. We describe the Riotee hardware consisting of a base module, a debug probe for easy firmware updates, and several expansion boards that enhance functionality without the need for custom-designed printed circuit boards (PCBs). The demo features the classification of live audio recordings using TinyML deep neural network inference aboard a Riotee device. The Riotee device transmits the classification results via Bluetooth Low Energy (BLE) to a smartphone given to visitors. Visitors can also observe how the Riotee software checkpoints and restores critical state in case of power failures via the visualization of logic analyzer traces.
- Open Access
- 4 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 198-210
Riotee: An Open-source Hardware and Software Platform for the Battery-free Internet of Things
The rapidly growing Internet of Things (IoT) can avoid the high cost and environmental burden of replacing trillions of batteries by using sustainable battery-free devices that operate maintenance-free for decades. To develop battery-free IoT systems, researchers and makers require a common platform that is versatile, affordable, and easy to use. However, limited availability and lack of support have prevented widespread adoption of previous battery-free platforms. We introduce Riotee, an open-source and commercially available battery-free platform that includes multiple boards, extensive software, and comprehensive documentation. We demonstrate Riotee's capabilities through a machine-learning application and present results from a user study involving students and customers, who rated its usefulness and usability highly.
- 4 November 2024
- Published by Association for Computing Machinery (ACM)
- p. 198-210
Open AccessRiotee: An Open-source Hardware and Software Platform for the Battery-free Internet of Things
The rapidly growing Internet of Things (IoT) can avoid the high cost and environmental burden of replacing trillions of batteries by using sustainable battery-free devices that operate maintenance-free for decades. To develop battery-free IoT systems, researchers and makers require a common platform that is versatile, affordable, and easy to use. However, limited availability and lack of support have prevented widespread adoption of previous battery-free platforms. We introduce Riotee, an open-source and commercially available battery-free platform that includes multiple boards, extensive software, and comprehensive documentation. We demonstrate Riotee's capabilities through a machine-learning application and present results from a user study involving students and customers, who rated its usefulness and usability highly.
- Open Access
- 9 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 354-355
Demo Abstract: Building Battery-free Devices with Riotee✱ | Proceedings of the 22nd International Conference on Information Processing in Sensor Networks
Battery-free devices eliminate the need for batteries, which are expensive, environmentally harmful, and require frequent replacement, thus reducing waste and making devices more cost-effective. We introduce Riotee, the next-generation platform for the battery-free Internet of Things. The platform comprises a base module, a debug probe that allows to conveniently update the firmware on the base module, and a number of expansion boards that extend the capabilities of the platform without the need to design a custom printed circuit board (PCB). We provide a brief overview of Riotee, and describe a demo setup that showcases the key functionality and how to get started with the platform in less than three minutes.
- 9 May 2023
- Published by Association for Computing Machinery (ACM)
- p. 354-355
Open AccessDemo Abstract: Building Battery-free Devices with Riotee✱ | Proceedings of the 22nd International Conference on Information Processing in Sensor Networks
Battery-free devices eliminate the need for batteries, which are expensive, environmentally harmful, and require frequent replacement, thus reducing waste and making devices more cost-effective. We introduce Riotee, the next-generation platform for the battery-free Internet of Things. The platform comprises a base module, a debug probe that allows to conveniently update the firmware on the base module, and a number of expansion boards that extend the capabilities of the platform without the need to design a custom printed circuit board (PCB). We provide a brief overview of Riotee, and describe a demo setup that showcases the key functionality and how to get started with the platform in less than three minutes.
- Open Access
- 25 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 23 (11) , 12411-12426
FusedAR: Energy-Positive Human Activity Recognition Using Kinetic and Solar Signal Fusion
Today’s wearable Internet of Things (IoT) devices, which have been fêted for numerous applications, suffer from the limited lifetime of batteries due to the high power consumption of conventional inertial activity sensors. Recently, kinetic energy harvesters have been employed as a source of energy as well as context information to replace conventional activity sensors. However, the harvested power from human movements using miniaturised kinetic transducers may not be sufficient to enable a perpetual and self-powered activity recognition system. In this paper, we propose a novel mechanism of Fused signal based human Activity Recognition (FusedAR), which employs miniaturised wearable solar and kinetic energy harvesters simultaneously as an energy source as well as an activity sensor. As human activities engender distinct movement patterns, and interact and interfere with the ambient light differently, the kinetic and solar energy harvesting signals incorporate unique information about the underlying activities while generating sufficient power. After detailed experiments, we find that FusedAR which employs both solar and kinetic energy signals, achieves superior activity recognition performance by up to 10%, particularly in outdoor and night-time contexts, and can recognise not only activities but also contexts, compared to the individual energy harvesting signals. Furthermore, our analysis demonstrates that FusedAR, in addition to significant energy generation, consumes up to 22% less power than the highly optimised conventional 3-axis accelerometer-based mechanisms, achieving energy-positive human activity recognition leading towards perpetual, uninterrupted and autonomous operation of wearable IoT devices.
- 25 April 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Sensors Journal
- Vol. 23 (11) , 12411-12426
Open AccessFusedAR: Energy-Positive Human Activity Recognition Using Kinetic and Solar Signal Fusion
Today’s wearable Internet of Things (IoT) devices, which have been fêted for numerous applications, suffer from the limited lifetime of batteries due to the high power consumption of conventional inertial activity sensors. Recently, kinetic energy harvesters have been employed as a source of energy as well as context information to replace conventional activity sensors. However, the harvested power from human movements using miniaturised kinetic transducers may not be sufficient to enable a perpetual and self-powered activity recognition system. In this paper, we propose a novel mechanism of Fused signal based human Activity Recognition (FusedAR), which employs miniaturised wearable solar and kinetic energy harvesters simultaneously as an energy source as well as an activity sensor. As human activities engender distinct movement patterns, and interact and interfere with the ambient light differently, the kinetic and solar energy harvesting signals incorporate unique information about the underlying activities while generating sufficient power. After detailed experiments, we find that FusedAR which employs both solar and kinetic energy signals, achieves superior activity recognition performance by up to 10%, particularly in outdoor and night-time contexts, and can recognise not only activities but also contexts, compared to the individual energy harvesting signals. Furthermore, our analysis demonstrates that FusedAR, in addition to significant energy generation, consumes up to 22% less power than the highly optimised conventional 3-axis accelerometer-based mechanisms, achieving energy-positive human activity recognition leading towards perpetual, uninterrupted and autonomous operation of wearable IoT devices.
- Open Access
- 22 January 2021
- Published by Association for Computing Machinery (ACM) in GetMobile: Mobile Computing and Communications
- Vol. 24 (3) , 5-8
Taking a Deep Dive Into The Batteryless Internet of Things With Shepherd
Collaboration of batteryless devices is essential to their success in replacing traditional battery-based systems. Without significant energy storage, spatio-temporal fluctuations of ambient energy availability become critical for the correct functioning of these systems. We present Shepherd, a testbed for the batteryless Internet of Things (IoT) that can record and reproduce spatio-temporal characteristics of real energy environments to obtain insights into the challenges and opportunities of operating groups of batteryless sensor nodes.
- 22 January 2021
- Published by Association for Computing Machinery (ACM) in GetMobile: Mobile Computing and Communications
- Vol. 24 (3) , 5-8
Open AccessTaking a Deep Dive Into The Batteryless Internet of Things With Shepherd
Collaboration of batteryless devices is essential to their success in replacing traditional battery-based systems. Without significant energy storage, spatio-temporal fluctuations of ambient energy availability become critical for the correct functioning of these systems. We present Shepherd, a testbed for the batteryless Internet of Things (IoT) that can record and reproduce spatio-temporal characteristics of real energy environments to obtain insights into the challenges and opportunities of operating groups of batteryless sensor nodes.