Battery, processor, GPS, and collision models

Battery

The approximate battery method uses a numerical matrix exponential and is the faster choice for repeated evaluation:

p_fail, mttf = model.Battery_Failure_Risk_Calc(
    BatteryLevel=80,
    time=10,
    Lambda=0.001,
    alpha=0.008,
    beta=0.007,
    Battery_degradation_rate=0.0064,
)

Battery_Failure_Risk_Calc_precise evaluates the corresponding symbolic model and can take longer.

Input	Meaning
BatteryLevel	Charge percentage from 0 to 100
Lambda	Battery-system component failure rate
alpha / beta	Charge and discharge transition rates
Battery_degradation_rate	Rate of transition to a degraded charge state

At 25% charge or below, the implementation reports failure with zero MTTF.

Processor

The processor model applies an Arrhenius acceleration factor to a reference MTTF, adjusts for utilization, and evaluates failure using a Weibull model.

p_fail, mttf = model.Chip_MTTF_Model(
    MTTFref=400,
    Tr=303.15,
    Ta=323.15,
    u=1,
    b=1,
    time=10,
)

Use absolute temperature consistently when applying an Arrhenius relationship.

GPS

The GPS method models the connection as a k-out-of-n system:

p_fail, mttf = model.GPS_Failure_Risk_Calc(
    SatStatus=22,
    time=10,
    Lambda=0.001,
    MaxSat=29,
    MinSat=17,
)

If fewer than MinSat satellites are available, the implementation returns failure with zero MTTF.

Collision and danger zones

calculate_collision_risk compares synchronized samples from two trajectories. It returns the fraction of samples inside the danger and collision thresholds:

uav_1 = [(0, 0), (1, 1), (2, 2)]
uav_2 = [(5, 5), (1.5, 1.5), (2.1, 2.1)]

danger_risk, collision_risk = model.calculate_collision_risk(
    uav_1,
    uav_2,
    danger_threshold=2.0,
    collision_threshold=0.5,
)

Both trajectories must contain the same number of samples and use the same coordinate system and distance unit as the thresholds.