2026-06-24T06:42:57Z https://api.figshare.com/v2/oai

oai:figshare.com:article/28738931 2025-06-03T10:33:16Z category_29161 category_29164 portal_549 item_type_3 month_year_06_2025

10.5522/04/28738931.v1 https://figshare.com/articles/dataset/CYCLEGYRO_Dataset/28738931 https://ndownloader.figshare.com/files/53454284 Cheng, None Cheng Deep learning Neural networks gyroscope sensor data Time Series Classification (TCS) action detection cycling data 2025-06-03 2025-06-03 Dataset 2025 University College London This dataset consists of multivariate time series data featuring gyroscope measurements collected during cycling. It contains 414 data samples. The original data was collected as part of UCL’s 100 Cyclists Project which seeks to understand the factors affecting crash risk in cycling through the analysis of detailed near miss data. In total, 64 participants were recruited, who used a GoPro Max panoramic camera mounted on their helmet to collect video and movement data while riding in London. The CYCLEGYRO dataset was built by extracting gyroscope clips from the IMU of the GoPro camera. These gyroscope clips contain 199 positive (there is one or more head movement) and 215 negative (there is no head movement) samples. Each sample is basically a three dimentional signal.

]]>