Measuring motion with kinematically redundant accelerometer arrays: Theory, simulation and implementation

Abstract

This work presents two schemes of measuring the linear and angular kinematics of a rigid body using a kinematically redundant array of triple-axis accelerometers with potential applications in biomechanics. A novel angular velocity estimation algorithm is proposed and evaluated that can compensate for angular velocity errors using measurements of the direction of gravity. Analysis and discussion of optimal sensor array characteristics are provided. A damped 2 axis pendulum was used to excite all 6 DoF of the a suspended accelerometer array through determined complex motion and is the basis of both simulation and experimental studies. The relationship between accuracy and sensor redundancy is investigated for arrays of up to 100 triple axis (300 accelerometer axes) accelerometers in simulation and 10 equivalent sensors (30 accelerometer axes) in the laboratory test rig. The paper also reports on the sensor calibration techniques and hardware implementation.

A Comparison of Day-Long Recording Stability and Muscle Force Prediction between BSN-Based Mechanomyography and Electromyography

“…At the moment, there are no economically viable solutions that enable large-scale collection of such data outside of laboratory or medical setups and in parallel offer sufficient data quality to be useful for clinical studies. This is because conventional motion and neurological signal capture equipment [15] [16] [17] [18] are usually bulky as they are designed for non-mobile laboratory/clinical use and often require external hardware which often restrict subjects in their daily-activities. A few kinematic BSN platforms have been designed in response to those problems [19] [20] [21], however only few studies have quantitatively evaluated their sensors performance and accuracy [18] [22] against a groundtruth, something that makes it hard to judge the suitability of the data collected in research and diagnostic applications.…”

Signal Quality and Compactness of a Dual-Accelerometer System for Gyro-Free Human Motion Analysis

“…Accelerometers measure linear acceleration with an associated measurement noise that can be modelled as an additive zero mean Gaussian noise [15]. The measurements on the i th 2D accelerometer on the body segment are modelled by a statistical observation model:…”

“…where R ∈ R 2×2 is the covariance matrix of the Gaussian noise ν t [15]. If the noise on both axes have the same statistical properties and are independent as in [15], then the covariance matrix simplifies as R = σ 2 I 2 with I 2 the identity matrix of R 2×2 and σ ∈ R + the standard deviation. It is reasonable to assume noise is independent as each accelerometer creates its own thermal noise.…”

“…Two 2D accelerometers are located on the body segment at 1 and 2 from the joint. Assuming the noise on both accelerometers is independent [15], the difference between two accelerometer signals is defined as:…”

“…The angular velocity is recoverable from either the angular acceleration term or the squared velocity term. The acceleration term can be integrated [5], alternatively the squared angular velocity term can be used as a computation of the absolute value of the velocity [2], with the sign estimated from other sources [15]. A pair or an array of accelerometers can also be used to estimate angle without integration in order to remove the body linear acceleration [3], [15].…”