Can gait analysis be performed in a web browser without installing an app?

How a browser reaches the phone’s motion sensors

Modern mobile browsers expose the phone’s motion sensors to web pages through standard device-motion and device-orientation interfaces. A web application can subscribe to those streams and receive acceleration and rotation readings many times a second, without any software being installed.

The data arrives as a time series of numbers in the page. From there it can be processed in the same way as data from any other inertial sensor: filtered, scanned for gait events, and turned into gait measures.

Permissions and secure context

• Sensor access requires a secure HTTPS connection. Browsers block these interfaces on unencrypted pages.
• On some platforms the user must explicitly grant motion-sensor permission, triggered by a deliberate action such as tapping a start button.
• Camera-based assessment needs a separate camera permission, requested only when that assessment is used.
• Permission is scoped to the site and the session, and the user can decline or revoke it at any time.

Good assessment design makes this explicit: tell the person what is being requested, why, and what happens to the data, before the permission prompt appears.

Calibration and setup

Because the browser receives raw sensor output, a short setup step does most of the work in making the resulting measures comparable.

1. 1Confirm the sensor stream is actually delivering data at a usable rate before the assessment starts.
2. 2Establish the phone’s resting orientation so movement can be interpreted relative to the body rather than the device.
3. 3Position the phone as the protocol specifies, and keep that placement consistent across repeat assessments.
4. 4Give clear instructions for the movement itself — where to walk, how far, at what effort.
5. 5Discard or flag a capture where the signal indicates the protocol was not followed.

Device and browser variability

This is the honest constraint of browser-delivered sensing. Phones differ in sensor hardware, and browsers differ in the sampling rate they deliver and how they behave when the screen dims or the page loses focus.

• Sampling rate can differ between browsers on the same handset.
• Some browsers throttle sensor delivery when the page is backgrounded.
• Sensor quality and noise characteristics vary between phone models and price points.
• Operating-system updates can change permission behaviour.

Checking the achieved sampling rate at capture time allows a service to identify and handle shortfalls across different devices and browsers.

Secure processing of the captured data

• Data is transmitted over encrypted connections only.
• Analysis runs server-side, so the phone does not need to be powerful.
• Results are returned to identified, authorised users rather than left on the device.
• Access to results follows the same access controls as any other clinical or study record.

Handling health-related data also brings information-governance obligations around lawful basis, retention and the person’s rights. Those are service-level decisions rather than a property of browser delivery itself.

Delivery by web link: SMS, email and QR codes

Once assessment runs in a browser, it can be delivered as a link. That changes the logistics considerably.

• SMS — a text message containing a secure link, opened on the phone the person already uses.
• Email — the same link where email suits the population better, or where a longer explanation is helpful.
• QR code — displayed in clinic or printed on an information sheet, so the person starts on their own device in the room.

Links should be single-purpose, time-limited and tied to the intended person, so that opening one is not equivalent to holding a general credential.

Why browser delivery reduces onboarding friction

1. 1No app store account, download, install or update is required.
2. 2No storage space is consumed on the person’s phone.
3. 3The person taps a link and starts, which removes the step where most drop-out happens.
4. 4Updates reach everyone immediately, so there is no fragmentation across app versions.
5. 5Nothing has to be posted, fitted, charged, collected or cleaned.

For older or less digitally confident populations, removing the install step is often the difference between an assessment being completed and being abandoned.

What MoveLab’s published study used

The published MoveLab validation study used a browser-based React data collection application accessing the smartphone’s built-in inertial sensors at a sampling rate of at least 50 Hz. That demonstrates browser-delivered capture was sufficient for the analysis performed in that study.

The study used a Samsung Galaxy A25 secured in a pouch at the waist with 25 healthy adults in a controlled laboratory environment.

How MoveLab is arranged

• The MoveLab Engine performs the analysis — turning captured sensor and camera data into movement measures.
• The MoveLab Platform is the ready-to-use software that uses that technology, including the assessment workflow and clinician-facing results.
• Partners can integrate the Engine into a product they already have, keeping their own patient experience.
• MoveLab can also host the browser-based assessment workflow, so a partner does not have to build the capture experience at all.