What comes next

Workstream one

Prototype 2: working without the app In development

The first prototype required a smartphone app to control all the cueing features. While the app is powerful and gives users a huge amount of flexibility, feedback from Parkinson's patients made one thing clear: not everyone wants to rely on a phone, and in the middle of a freezing episode, reaching for a screen isn't always realistic.

Prototype 2 moves the controls onto the stick itself.

By migrating from the Raspberry Pi Zero 2W to an ESP32 microcontroller, the device becomes smaller, lighter and cheaper. Physical buttons built directly into the walking stick handle will allow users to toggle the audio beat, vibration and laser on and off without needing their phone at all. The smartphone app will still be available for users who want fine-grained control and session recording, but the device will work completely independently without it.

The custom PCB design developed in Prototype 1 will be refined and adapted for the ESP32, making the hardware more compact and robust.

Currently the device relies on the user manually activating the cueing when they feel a freeze coming. The vision for the next stage of the project is for the device to detect a Freezing of Gait episode automatically and respond without any input needed from the user.

The foundations are already in place. The MPU6050 motion sensor built into the stick continuously captures accelerometer and gyroscope data, and the companion app can already record and export walking sessions. By collecting labelled gait data, walking normally and walking with freezing episodes, a machine learning model can be trained to recognise the characteristic signal patterns of a freeze in real time.

The genuinely hard problem, and the heart of the research, is distinguishing a real freezing episode from an intentional stop. A person pausing at a kerb, turning to talk to someone, or simply choosing to stand still produces sensor data that can look superficially similar to a freeze. A useful detector has to tell these situations apart reliably, because a device that cues every time its user chooses to stop would quickly become a device its user turns off.

Once a freeze is detected, the device would be able to propose or automatically trigger the appropriate cueing response, audio, haptic or visual, to help the user break the freeze as quickly as possible, even in cases where they are unable to interact with the device manually.

This would bring the Smart Walking Stick considerably closer to a truly autonomous assistive device, and represent a meaningful step forward in what technology can offer Parkinson's patients.

The bigger picture

Where this is all heading

The long-term goal is straightforward: to turn the Smart Walking Stick into a clinically validated, affordable assistive device that can reach the people who need it.

Existing multimodal cueing devices are expensive and inaccessible for many patients. The Smart Walking Stick is being built from the ground up with accessibility and affordability as core design principles: accessible hardware, open development and a pragmatic approach to cost.

There is a real gap in the market and a real clinical need. The project has already demonstrated that people with Parkinson's respond positively to the device. The next steps are to build the evidence, improve the hardware, and find a path to getting it into the hands of patients, whether through the NHS, Parkinson's charities, or directly to the people who need it.