Detection of Wheelchair User Activities Using Wearable Sensors

Hiremath, Shivayogi V.; Chung, Younghyun; Cooper, Rory A.

doi:10.1007/978-3-642-21666-4_17

Cited by 13 publications

(8 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…However, the fact that other types of interaction (e.g. ADLs [5] or wheelchair stroke patterns [6]) can be detected using Human Activity Recognition methods suggests that a semi-automated approach could be effective, by detecting where a user might be encountering a problem and then inviting an annotation from the user.…”

Section: Automated (Or Semi-automated)mentioning

confidence: 99%

WheelieMap

Kirkham

Ebassa

Montague

et al. 2017

Proceedings of the 19th International Conference on Human-Computer Interaction With Mobile Devices and Services

View full text Add to dashboard Cite

The built environment remains a persistent accessibility challenge for people with mobility impairments. Whilst platforms to report these inaccessible locations exist, the underlying documentation processes are verbose, timeconsuming and fail to effectively communicate the barrier at hand. We propose WheelieMap, a platform which uses the motion of manual wheelchair users to support the identification and documentation of potentially problematic locations. WheelieMap captures and segments device video footage and GPS as evidence of the problematic space, which can then be shared with both other people with disabilities and the relevant authorities. We document the use of the WheelieMap prototype by both manual wheelchair users and planning experts through semi-structured interviews. The qualitative findings revealed this approach to be the most viable route for documenting inaccessibility, compared to the existing alternatives. We also offer guidance on how to design and develop similar community driven reporting and annotation systems in the accessibility setting.

show abstract

Section: Automated (Or Semi-automated)mentioning

confidence: 99%

WheelieMap

Kirkham

Ebassa

Montague

et al. 2017

Proceedings of the 19th International Conference on Human-Computer Interaction With Mobile Devices and Services

View full text Add to dashboard Cite

show abstract

“…Wearable sensors such as inertial measurement units (IMUs) and electromyography (EMG) sensors are minimally intrusive measurement tools that can be used to quantify movements [ 37 , 38 ] and muscle activity over a longer period. The largest limitation of using IMUs lies in “integration drift” when fusing sensor signals into orientation estimates [ 39 ] and the temperature depending bias of gyroscopes has been reduced to a minimum in the more recent generations of sensors.…”

Section: Introductionmentioning

confidence: 99%

“…Therefore, the use of only raw (or calibrated) wearable sensor data such as the acceleration and angular velocity for the classification of activities is a growing field of research, with promising results. Especially in the able-bodied population, IMUs or embedded sensors (e.g., smartphones, smartwatches) have been shown to be a valuable tool to monitor activities in a free-living environment [ 40 , 41 , 42 ], but only a few studies have investigated activity detection and classification among MWUs [ 37 , 43 , 44 , 45 , 46 , 47 , 48 , 49 ]. Previous research examining the use of wearable sensors has predominantly focused on physical activity detection to estimate the activity levels and energy expenditure in MWUs with SCI [ 50 , 51 , 52 , 53 , 54 , 55 ].…”

Section: Introductionmentioning

confidence: 99%

Classification of Wheelchair Related Shoulder Loading Activities from Wearable Sensor Data: A Machine Learning Approach

Vries

Amrein

Arnet

et al. 2022

Sensors

View full text Add to dashboard Cite

Shoulder problems (pain and pathology) are highly prevalent in manual wheelchair users with spinal cord injury. These problems lead to limitations in activities of daily life (ADL), labor- and leisure participation, and increase the health care costs. Shoulder problems are often associated with the long-term reliance on the upper limbs, and the accompanying “shoulder load”. To make an estimation of daily shoulder load, it is crucial to know which ADL are performed and how these are executed in the free-living environment (in terms of magnitude, frequency, and duration). The aim of this study was to develop and validate methodology for the classification of wheelchair related shoulder loading ADL (SL-ADL) from wearable sensor data. Ten able bodied participants equipped with five Shimmer sensors on a wheelchair and upper extremity performed eight relevant SL-ADL. Deep learning networks using bidirectional long short-term memory networks were trained on sensor data (acceleration, gyroscope signals and EMG), using video annotated activities as the target. Overall, the trained algorithm performed well, with an accuracy of 98% and specificity of 99%. When reducing the input for training the network to data from only one sensor, the overall performance decreased to around 80% for all performance measures. The use of only forearm sensor data led to a better performance than the use of the upper arm sensor data. It can be concluded that a generalizable algorithm could be trained by a deep learning network to classify wheelchair related SL-ADL from the wearable sensor data.

show abstract

“…Also using the wheel rotation monitor to identify self-propulsion episodes may not be accurate in real-life settings. Our previous study showed that the monitors we used here were able to detect self-propulsion, external pushing, sedentary activities, and other activities with an accuracy of 90% using a laboratory-based protocol [ 37 ]. Future testing should consider real-life testing with a mixture of wheelchair propulsion and other activities of daily living in the home and community settings and combine the detection of wheelchair episodes with the estimation of propulsion parameters when assessing the overall estimation accuracies of temporal propulsion parameters.…”

Section: Discussionmentioning

confidence: 99%

Temporal Parameters Estimation for Wheelchair Propulsion Using Wearable Sensors

Ojeda

2014

BioMed Research International

Self Cite

View full text Add to dashboard Cite

Due to lower limb paralysis, individuals with spinal cord injury (SCI) rely on their upper limbs for mobility. The prevalence of upper extremity pain and injury is high among this population. We evaluated the performance of three triaxis accelerometers placed on the upper arm, wrist, and under the wheelchair, to estimate temporal parameters of wheelchair propulsion. Twenty-six participants with SCI were asked to push their wheelchair equipped with a SMARTWheel. The estimated stroke number was compared with the criterion from video observations and the estimated push frequency was compared with the criterion from the SMARTWheel. Mean absolute errors (MAE) and mean absolute percentage of error (MAPE) were calculated. Intraclass correlation coefficients and Bland-Altman plots were used to assess the agreement. Results showed reasonable accuracies especially using the accelerometer placed on the upper arm where the MAPE was 8.0% for stroke number and 12.9% for push frequency. The ICC was 0.994 for stroke number and 0.916 for push frequency. The wrist and seat accelerometer showed lower accuracy with a MAPE for the stroke number of 10.8% and 13.4% and ICC of 0.990 and 0.984, respectively. Results suggested that accelerometers could be an option for monitoring temporal parameters of wheelchair propulsion.

show abstract

Detection of Wheelchair User Activities Using Wearable Sensors

Cited by 13 publications

References 14 publications

WheelieMap

WheelieMap

Classification of Wheelchair Related Shoulder Loading Activities from Wearable Sensor Data: A Machine Learning Approach

Temporal Parameters Estimation for Wheelchair Propulsion Using Wearable Sensors

Contact Info

Product

Resources

About