Reliability models

SafeDrones combines several subsystem models behind one Python class. Each model accepts the latest observed state and returns a failure probability and an MTTF estimate.

Model map

Model	Main inputs	Method
Propulsion	Motor status, rotor configuration, failure rate, mission time	Motor_Failure_Risk_Calc
Battery	Charge level, failure/degradation rates, charge/discharge rates, time	Battery_Failure_Risk_Calc
Processor	Reference MTTF, reference/actual temperature, utilization, Weibull beta	Chip_MTTF_Model
GPS	Visible satellites, required satellites, single-link failure rate, time	GPS_Failure_Risk_Calc
Collision	Two sampled trajectories, danger and collision thresholds	calculate_collision_risk
Combined drone	Current configured propulsion, battery, and processor state	Drone_Risk_Calc

Combined risk

Drone_Risk_Calc treats propulsion, battery, and processor failure as independent for its combined probability:

P(total failure) = 1 - (1 - Pmotor)(1 - Pbattery)(1 - Pprocessor)

The combined MTTF is the minimum of the three subsystem estimates.

Model assumptions matter

Independence and constant-rate assumptions are modelling choices, not guarantees about a physical aircraft. Validate input rates and thresholds against the vehicle, operational environment, and applicable assurance process before using results in decisions.

Selecting a model

• Use the propulsion model when rotor placement and individual motor state determine controllability after failures.
• Use the component models for battery, processor, or GPS reliability.
• Use combined drone risk for a compact mission-level indicator after all input values have been configured consistently.