Analyzing human gait patterns for malfunction detection

Köhle, Monika; Merkl, Dieter

doi:10.1145/335603.335629

Cited by 20 publications

(12 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, disappointing results led to further investigation. Kohle and Merkl [5] suggested that the number of degrees of freedom is a direct cause of the variability in walking and therefore, unrestricted natural walking is likely to show the most variable profile in contrast to someone with a walking disorder. Given that affected gait is usually down to a limitation in limb movement, be it from muscular or nervous defects, it was justifiable to classify just these limited movements, rather than the many classes of 'natural' walking.…”

Section: Distance Calculationmentioning

confidence: 99%

A smart sensing platform for the classification of ambulatory patterns

Elliott¹,

Ma²,

Brett³

2009

Proc Inst Mech Eng H

View full text Add to dashboard Cite

This paper describes an innovative sensing approach allowing capture, discrimination, and classification of transients automatically in gait. A walking platform is described, which offers an alternative design to that of standard force plates with advantages that include mechanical simplicity and less restriction on dimensions. The scope of the work is to investigate as an experiment the sensitivity of the distributive tactile sensing method with the potential to address flexibility on gait assessment, including patient targeting and the extension to a variety of ambulatory applications. Using infrared sensors to measure plate deflection, gait patterns are compared with stored templates using a pattern recognition algorithm. This information is input into a neural network to classify normal and affected walking events, with a classification accuracy of just under 90 per cent achieved. The system developed has potential applications in gait analysis and rehabilitation, whereby it can be used as a tool for early diagnosis of walking disorders or to determine changes between pre- and post-operative gait.

show abstract

Section: Distance Calculationmentioning

confidence: 99%

A smart sensing platform for the classification of ambulatory patterns

Elliott¹,

Ma²,

Brett³

2009

Proc Inst Mech Eng H

View full text Add to dashboard Cite

show abstract

“…The lowfrequency (below 500 Hz) responses are well known from the previous investigations. [1][2][3][4][5][6] The vibration signal above 500 Hz (Figs. 2(a) and 2(b)) has a broadband frequency spectrum up to 22 kHz with high-frequency magnitudes comparable with the maxima of magnitudes at the low frequencies.…”

Section: Footsteps Vibration and Sound Signatures In A Buildingmentioning

confidence: 99%

“…This force is due to two components: the normal force and the force tangential to the surface. 6 The normal force component generates the low-frequency vibrations (typically below 500 Hz) that are currently used for intruder detection by seismic methods. 5 The tangential force is governed by the horizontal motion and is equal to the force of the ground/floor reaction, or friction force.…”

Section: Introductionmentioning

confidence: 99%

Broad frequency acoustic response of ground/floor to human footsteps

Ekimov

Sabatier

2006

SPIE Proceedings

View full text Add to dashboard Cite

The human footstep is one of several signatures that can serve as a useful parameter for human detection. In early research, the force of footsteps was measured on load cells and the input energy from multiple footsteps was detected in the frequency range of 1-4 Hz. Cress investigated the seismic velocity response of outdoor ground sites to individuals that were crawling, walking, and running. In his work, the seismic velocity response was shown to be site-specific and the characteristic frequency range was 20-90 Hz. The current paper will present vibration and sound pressure responses of human footsteps in a broad frequency range. The vibration and sound in the low-frequency band are well known in the literature and generated by the force component normal to the ground/floor. This force is a function of person's weight and a manner of motion (e.g. walking, running, etc). Forces tangential to the ground/floor from a footstep and the ground reaction generate the high frequency responses. The interactions of foot and the ground/floor produce sliding contacts and the result is a friction signal. The parameters of this friction signal, such as frequency band and vibration and sound magnitudes as functions of human walking styles, were studied. The results of tests are presented and discussed.

show abstract

“…Traditional such research is conducted using dynamic electromyography, force platforms and joint kinematics and kinetics. [4], [5], [9] In our study, we perform a computing project based on the tri-axial accelerometers that capture human movement patterns from various body locations. The goal ofthis study is to classifying different age groups based on this movement information.…”

Section: Introductionmentioning

confidence: 99%

“…Kohle and Merkl [4] classified a number of pathological gait patterns by using the unsupervised learning strategy of self-organizing maps (SOM). Yun, Woo and Ryu [9] used a neural network that was trained using the feedback from the photo interrupter sensors that were uniformly arrange in each floor tiles.…”

Section: Introductionmentioning

confidence: 99%

Mining from Time Series Human Movement Data

Tseng

Cook

2006

2006 IEEE International Conference on Systems, Man and Cybernetics

View full text Add to dashboard Cite

Human motion not only contains a wealth of information about actions and intentions, but also about identity and personal attributes of the moving person. Research also indicates that there are positive relationships between the health of a person and their pattern of motion. In this research we utilize human walking data collected from volunteers to identify age categories and to detect possible changes in the individual's health condition.The approach is based on transforming biological motion data into a representation that subsequently allows for analysis using artificial intelligence techniques. Using wireless accelerometer sensors we were able to collect and transform the data into a numeric representation. We then applied the numeric data to various artificial intelligence algorithms to form classification models and use it for our analysis.

show abstract

Analyzing human gait patterns for malfunction detection

Cited by 20 publications

References 7 publications

A smart sensing platform for the classification of ambulatory patterns

A smart sensing platform for the classification of ambulatory patterns

Broad frequency acoustic response of ground/floor to human footsteps

Mining from Time Series Human Movement Data

Contact Info

Product

Resources

About