Surface Type Classification for Autonomous Robot Indoor Navigation - Dataset

Surface Type Recognition with Inertial Measurement Unit (IMU).

The dataset contains time series samples with 10 features each, related to orientation, velocity and acceleration. Each time series (of lenght 128) includes its corresponding surface type annotation.

The data has been also divided in groups for easier cross-validation (80 groups present)

A total of 9 different surface types are present in the dataset.

"X_data.npy" contains the time series samples of dimension 7626x10x128
"label.npy" contains the label information for each sample (dimension 7626x1)
"groups.npy" contains the group information for each sample (dimension 7626x1)
"details.csv" contains for each sample the group information and the corresponding label