Movement pattern analysis is an established research field. We make transparent what our approach builds on — and invite scientists and practitioners to expand the collection with us.
The most defensible scientific positioning is not "we diagnose injuries from two foot sensors" but: we detect clinically and athletically relevant deviations from the individual movement baseline, quantify risk and recovery signatures, and support medical decisions longitudinally. That is exactly what today’s literature robustly supports — and it does not preclude tightly validated injury modules later.
Gait speed, cadence, step/stride length, step time and variability, double-support ratio, step width, turning metrics and left-right asymmetry.
Optical marker systems remain the reference standard; IMU wearables provide real-world reach; markerless video closes the gap for spatiotemporal parameters.
Research is shifting from one-off tests to thresholds for clinically meaningful change over time — exactly our product’s logic.
Not a claim to a complete systematic review, but the decision-relevant evidence: robust methods to measure movement change. Legend: Peer-reviewed Preprint Project/report
Frame the feasibility window: what do which systems measure, and why do standardised parameter definitions matter?
Sy, N. et al. — Wearable inertial sensors for gait analysis: a technology overview Peer-Reviewed
Describes IMUs (accelerometer + gyroscope + magnetometer), derivation of orientation/position and the core trade-off "few sensors vs. high biomechanical resolution". Base source for realistically framing technical feasibility.
Review: standardised parameter definitions & validation in movement analysis Peer-Reviewed
Emphasises that changes only remain robustly comparable across time points, systems and studies if parameters are standardised and validated — a core requirement for our data ontology.
Methodologically cleanest: optoelectronic multi-camera systems plus force plates as the gold standard.
Jerome, G. J. et al. — Gait characteristics associated with walking speed decline in older adults (Baltimore Longitudinal Study of Aging) Peer-Reviewed
3D motion capture + force plates, multi-year follow-up: 23% showed a clinically relevant speed decline over ~3.2 years; longer strides and higher cadence were associated with lower risk. Proves that simple spatiotemporal metrics carry prospective, longitudinal relevance.
Biomechanical analysis of gait patterns in healthy individuals Peer-Reviewed
Measures gait cycles three-dimensionally with optoelectronic systems and force plates — provides the healthy-movement reference values against which deviations are measured.
The strongest bridge between lab and product: wearable IMUs capture movement in daily life over days to years.
Real-life gait assessment in degenerative cerebellar ataxia: 2-year longitudinal study with body-worn IMUs Peer-Reviewed
Three body-worn IMUs under real walking conditions: variability measures (compound measure SPCmp) detect significant change after just one year — before clinical scales respond. High test-retest reliability (ICC ≥ 0.82).
Gait Variability as a Potential Motor Marker of Cerebellar Disease Peer-Reviewed
Investigates the relationship between variability of stride, arm swing and trunk movement and walking speed — shows which variability measures work as early markers.
Raykov, Y. P. et al. — Probabilistic modelling of free-living gait segments (Parkinson) Preprint
Addresses the real everyday problem: robust localisation of meaningful walking segments in unstructured data (sensor positions wrist/ankle/pocket/lower back). Central for a product outside standardised lab tests. As a preprint, to be weighted more cautiously.
Rinderknecht, M. D. et al. — MCIDs of digital gait measures in progressive multiple sclerosis Preprint
Daily smartphone-based two-minute walk tests; derives minimal clinically important differences (MCIDs) for step velocity, duration and length. Shows the direction: away from the test date, toward thresholds for meaningful change. Preprint.
Pose estimation from 2D/3D video, compared against the optoelectronic gold standard — scalable and accessible.
Molteni, L. & Andreoni, G. — Validation of OpenPose against an optoelectronic system for lower-limb gait Peer-Reviewed
Spatiotemporal parameters with good-to-excellent ICCs; small absolute errors for cadence and mean speed; good agreement of range of motion in the sagittal plane. Evidence that markerless video is realistic for scalable spatiotemporal analytics.
Evaluation of markerless motion capture against gold-standard systems Peer-Reviewed
Compares pose-estimation frameworks with optoelectronics; reports good-to-excellent agreement for spatiotemporal gait parameters, particularly in the sagittal plane.
Donahue, S. et al. — Uncertainty-calibrated multiview markerless gait analysis Preprint
Median errors of ~12-16 mm for step/stride length, bias-corrected kinematic errors of ~1.5-3.8°; plus calibrated estimation uncertainty. Shows where the stack can grow: not just estimate values, but state how certain they are. Preprint.
Funded projects translating the "norm-plus-deviation" logic into care — the strongest template for our product.
KIT — movement analysis for early osteoarthritis detection Research project
Models a catalogue of normal movement patterns and describes deviations probabilistically: restricted knee angle and slowed weight transfer become mathematically detectable early, before they are visually obvious. Exactly our logic — personal/population norm plus early measurable deviation.
AktiSmart-KI — Sensor-based activity and movement pattern recognition German MoH funded
Body-worn sensors recognise activities and movement patterns, detect falls, document treatment progress and derive treatment needs — reported activity recognition up to 85%. Shows the immediate value lies in trajectory data, early warning and decision support. Basis for our "Sensors for Seniors" use case.
Coarse population-level activity profiles — the outer layer of the problem, complementing the biomechanical depth layer.
RKI — HBSC analysis: physical activity of children and adolescents Institute/report
Works with macroscopic activity profiles (activity guidelines, sports activity, trends since 2009/10). Useful for public health questions — but injury detection requires finer, biomechanically interpretable signatures.
This collection is a curated selection of decision-relevant evidence, not a formal systematic review. Preprints are not yet peer-reviewed and should be weighted cautiously. All links lead to the original third-party sources.
We expand this directory together with scientists, physicians, coaches and the community. Every suggestion is reviewed by us before it is added.