User:Ron.bur999/sandbox

The AX3 is an open-source data logger for acceleration and is used for human, animal and other movement measurement and data collection. AX3 is one of a suite of hardware devices created and maintained within the Open Movement initiative at Newcastle University, UK, and was developed in response to UK Biobank's requirement for a low-cost, high precision device to be used in its collection of 7-day data for 100,000 participants. Following its launch in 2012, and subsequent use by UK Biobank, the device rapidly became popular with researchers, given its relative low cost, ability to collect of raw actigraphy data, and the availability for comparison of UK Biobank data collected with the AX3. In 2015 a review and expert consensus study of the utilisation of adult actigraphy data identified that the AX3 had been used for over a third (38%) of the global pool of 277,370 adults for whom such data has been collected (more than any other device).

History
The AX3 was announced in October 2012 within a demonstration and presentation of Open Movement at Digital Futures, the EPSRC Digital Economy All Hands meeting in Salford, UK. Although the WAX3, a small form factor wireless accelerometer, used in the Ambient Kitchen and other IoT prototypes was the original Open Movement device, the AX3 was developed in response to the proposal for a large-scale physical activity data collection exercise by UK Biobank. The Open Movement initiative (which included the AX3) was established and led by Patrick Olivier, and involved Karim Ladha (hardware lead), Daniel Jackson (software lead), Cassim Ladha , Matt Kipling (original enclosure), Tom Nappey (enclosure, wristband and docking station) and Nils Hammerla (original algorithms). Open Movement is currently maintained and developed by Daniel Jackson (technical lead) at Newcastle University, and its AX3 and AX6 (6-axis logging accelerometer and gyroscope) are manufactured and distributed by Newcastle University spin-out Axivity Ltd.

Functional overview
The AX3 is used to measure movement, vibrations and orientation changes in 3-axes. In addition to having a configurable sampling rate, range and precision, it is distinct from similar devices in that it collects raw sensor data without forced sampling.[Ref] The AX3's triaxial MEMS accelerometer provides nine native sampling frequencies, from 12.5 Hz, doubling at each level to 3200 Hz, and ranges for acceleration of ±2/4/8/16g with 13-bit precision.

The rechargeable lithium polymer battery and internal non-volatile memory allow the recording of up to 30 days of continuous data (at 12.5 Hz), although in its default sampling rate for human motion (100 Hz) the device has battery and memory capacity for recording a 14-day dataset. Data is stored in raw format with built-in error checking and correction (ECC) and each sample is stored along with a precise time-stamp, allowing for accurate signal reconstruction (no truncation or lumping artefacts). When connected to a PC via the device's microUSB connector the AX3 can be re-charged, configured and its data downloaded using the open-source OMGUI software.

Additional sensors include a temperature sensor, which records the temperature at the surface of the device's PCB, and a light sensor that is tuned to visible wavelengths, and is also mounted inside the device enclosure. A single colour LED indicates the state of the sensor (i.e. charging, standby, recording or communicating to a host computer) but can also be configured in a “silent” logging mode for which the LED is turned off when logging. The device is CE marked and the enclosure is IP Code rated for moisture ingress water-resistant to 1.5m for 1 hour (IPx8) and dust tight (IP6x).

Adoption
The use of the AX3 by UK Biobank meant that in 2015 the AX3 had been used for over a third (38%) of the global pool of 277,370 adults for whom such data has been collected (more than any other device). However, this share of the global pool is now likely to be significantly higher given the subsequent use of the AX3 in a number of very large cohort studies, including version 4 of the Trøndelag Health Study (involving data collection from ~45,000 participants, each wearing two AX3 devices, one on the thigh and one on the back).

Equivalence
Multiple published studies have documented the degree equivalence between the AX3 with other devices historically used for human movement studies. This includes direct equivalence raw data recordings, but also equivalence of derived measures made by different devices (such as step and activity counts).

Validation
Correspondence of the AX3 with ground truth observations (e.g. step numbers) and biomarkers of energy expenditure (e.g. doubly labelled water) have been demonstrated in. A number of studies have also documented these for sub-populations, for example, for children, older adults, and those affected by neurodegenerative illness.

General Population (Physical Activity and Energy Expenditure)
Older Adults (Physical Activity and Energy Expenditure)

Children and Infants

Use cases
The AX3 was initially developed for lower the costs of large-scale actigraphy data collection in epidemiology, but is now used in a wide range of contexts, including intervention effectiveness (where physical activity is an outcome measure), biomechanical biomarker development, human-activity recognition development, and animal behaviour studies.

Epidemiology
Between February 2013 and December 2015, UK Biobank conducted the largest ever actigraphy data collection activity when it used 4043 AX3 devices to collect 7-day actigraphy from 103,712 members of its 500,000 participant cohort. With the goal of laying a "foundation for studies of physical activity and its health consequences" UK Biobank published all data, a preliminary analysis, and it analysis code.

Technology
As part of an open source project the hardware designs for the AX3, including PCB designs, layouts and schematics, as well as the 3D model of the enclosure design, are available under a Creative Commons 3.0 BY Attribution license. The firmware is available under a BSD 2-clause license.

Electronics
The primary components used in the AX3 are a non-volatile NAND flash memory chip, USB enabled microcontroller, 3-axis MEMS accelerometer (range 12.5-3200Hz, precision up to 13 bit), quartz real time clock (32.768KHz, ± 50ppm), logarithmic light sensor (470-650 nm, 3-1000 LUX, 10 bit), linear thermistor temperature sensor (range 0-40˚C, resolution 0.3˚C) and a rechargeable lithium polymer battery (capacity 150mAh, charge current 150mA).

Physical Design
The electronics is packaged within a sealed polycarbonate enclosure or "puck" (dimensions 23x32.5x8.9mm; total weight 11g) with a moisture (IPx8 water to 1.5m) and dust ingress resistant (IP6x) microUSB connector. For most human movement studies of physical activities the puck is used with the Open Movement skin-safe silicone material wrist band (weight 16g, adjustable 140-217mm).

Software
The OMGUI open-source software tool used to configure AX3 devices and to download and analyse sensor data post-deployment. The stable multi-device version of OMGUI is a Windows application (XP SP3 or later, Microsoft .NET Framework v3.5 or later) although there is an experimental web-based version for single devices that can be used with Mac OS, Linux and some other operating systems.

Configuration and Data Management
OMGUI allows users to configure devices for recording, including: clearing existing data, configuring devices to record (setting session ID, sampling frequency, sampling range, interval start/stop date/time delay/duration, study and subject meta data, and other parameters), as well as preview visualization of device data and downloading data. OMGUI can also be used to export device data in a range formats (raw CSV, re-sampled CSV, resample WAV) for visualization and analysis using third party applications.

Data Analysis
OMGUI incorporates a number of open source implementation of commonly used validated analysis algorithms that can be applied directly to data collected using the AX3. These include: (1) Signal Vector Magnitude; (2) Cut-point analysis (physical activity intensity bands in units of Metabolic Equivalent of Task) ; (3) Wear Time Validation (WTV) ; (4) Sleep, using the estimation of stationary sleep-segments approach. Increasingly, third party analysis algorithms for AX3 device collected data are made available by researchers when publishing results, for example, the analysis routines (for time spent in sleep, sedentary behaviour, and doing physical activity) applied to the data of 96,600 participants collected by UK Biobank.