Dataset Anomaly Detection in Underwater Drone Surveillance Systems

Each row: One time-step from a drone during surveillanceLabel: Normal or Anomalous (binary or multi-class depending on attack type or anomaly)

📊 Key Dataset Features (Grouped by Subsystem)

1. Navigation & Positional Features

Feature Name
Description

timestamp
Time of recording (UTC or local)

depth_m
Depth in meters

latitude
GPS latitude (if available in shallow water)

longitude
GPS longitude (same note as above)

velocity_mps
Linear velocity of the drone (m/s)

acceleration_mps2
Instantaneous acceleration

orientation_pitch
Drone pitch (degrees)

orientation_yaw
Drone yaw (degrees)

orientation_roll
Drone roll (degrees)

2. Sensor and Environmental Features

Feature Name
Description

water_temp_c
Water temperature in Celsius

salinity_psu
Salinity in PSU (Practical Salinity Units)

pressure_dbar
Water pressure at current depth (in decibars)

sonar_range_m
Distance to nearest object from sonar (in meters)

sonar_reflectivity
Strength of sonar signal reflection

camera_motion_blur
Blur score from image feed (to detect obstruction or anomalies)

video_object_count
Number of objects detected in current frame

turbidity_ntu
Water clarity (Nephelometric Turbidity Units)

magnetic_field
Magnetic field strength (microtesla) – detects magnetic anomalies

3. Operational/Control Features

Feature Name
Description

battery_voltage
Battery voltage level

motor_current
Electrical current drawn by motor (A)

thruster_force
Force applied by each thruster (N)

communication_delay
Latency in command transmission (ms)

packet_loss_rate
Loss rate of data packets during underwater transmission (%)

cpu_utilization
Percentage of onboard processor utilization

memory_usage
Onboard memory usage (%)

4. Anomaly Indicator (Label)

Feature Name
Description

anomaly_type
Categorical: e.g., "None", "Sonar Spoofing", "Sensor Failure", "Obstacle Detected", "Communication Jam"

anomaly_binary
0 = Normal, 1 = Anomaly