BVLOS and risk levels

Why this matters

UAT 312 explained what BVLOS is, radio horizon, C2 links, Detect and Avoid and Remote ID, and UAT 401 practised ground risk assessment with SORA. This practical course prepares BVLOS missions in earnest. The case used throughout is inspecting a 20 km raw-water pipeline across rural rice fields, an area where crop-spraying aircraft routinely fly low. The drone knowledge base’s unit on beyond-visual-line-of-sight operations covers DAA, observers, C2 and permission procedures, and its unit on BVLOS and risk assessment stresses linking risks, mitigations and evidence. The first module starts with the risk BVLOS raises most clearly: mid-air collision.

Initial ARC from airspace type

SORA 2.5 assesses air risk with the ARC (Air Risk Class). Annex C divides airspace into 12 Airspace Encounter Categories (AECs). For example, an airport environment in class B, C or D airspace gives ARC-d; above 500 ft AGL in controlled airspace gives ARC-d; above 500 ft in uncontrolled airspace gives ARC-c; below 500 ft in controlled airspace or over urban areas gives ARC-c; below 500 ft in uncontrolled airspace over rural areas gives ARC-b; and segregated or atypical airspace gives ARC-a. Annex C Table 2 also lets an applicant lower the ARC by showing that local traffic density is lower than assumed.

After strategic mitigation, the remaining (residual) ARC sets the TMPR under SORA 2.5 Table 6: ARC-d needs High, ARC-c Medium, ARC-b Low, and ARC-a has no requirement. Visual line of sight counts as an acceptable tactical mitigation at every level, so BVLOS must find something to replace the pilot’s eyes.

The traffic-density model

Weibel and Hansman (2004) estimated the mid-air collision risk of unmanned aircraft from the density of air traffic in each region, and found risk on major airways about a hundred times that off airways. The common underlying idea is the gas model, like counting how often gas molecules collide: encounter rate equals traffic density × the cross-section of the region counted as an encounter × relative speed. The result depends heavily on assumptions, so it is better for comparing options than as an absolute safety figure.

import math

# gas model: encounter rate = traffic density × cross-section × relative speed
aircraft = 0.5               # average crewed aircraft flying low in the area at once (assumed)
area_km2 = 10000             # area those aircraft use
band_km = 0.15               # height band 0–150 m
rho = aircraft / (area_km2 * band_km)          # aircraft per km³
r_km, h_km = 0.5, 0.1        # watch cylinder: radius 500 m, height ±100 m
cross = 2 * r_km * 2 * h_km  # km² cross-section facing the relative motion
v_rel = 60 * 3.6             # km/h (60 m/s)

rate = rho * cross * v_rel   # per flight hour
print(f"traffic density {rho:.2e} aircraft/km^3")
print(f"encounters inside the cylinder: {rate:.4f} per flight hour"
      f" = one per {1/rate:.0f} h")
hours = 300
p_any = 1 - math.exp(-rate * hours)
print(f"chance of at least one in {hours} flight hours: {p_any:.0%}")

traffic density 3.33e-04 aircraft/km^3
encounters inside the cylinder: 0.0144 per flight hour = one per 69 h
chance of at least one in 300 flight hours: 99%

Strategic mitigation by time of day

Strategic mitigation reduces encounters before take-off, for example by choosing heights, routes or times with little traffic. Talking with local crop-spraying operators may reveal that they fly most in the calm morning and evening.

import math

# average crop sprayers flying low at once, by hour (assumed from a survey of local operators)
by_hour = {6: 1.2, 7: 1.6, 8: 1.4, 9: 1.0, 10: 0.6, 11: 0.4, 12: 0.2, 13: 0.2, 14: 0.4, 15: 0.6, 16: 1.0, 17: 1.2}
area_km2, band_km = 10000, 0.15
cross = 2 * 0.5 * 2 * 0.1
v_rel = 60 * 3.6

def rate(n):
    return n / (area_km2 * band_km) * cross * v_rel

all_day = sum(rate(n) for n in by_hour.values()) / len(by_hour)
window = [11, 12, 13]
limited = sum(rate(by_hour[h]) for h in window) / len(window)
print(f"average over 06-18 h: {all_day:.4f} encounters per flight hour")
print(f"flying only 11-14 h:  {limited:.4f} encounters per flight hour")
print(f"reduction factor {all_day/limited:.1f}x")

average over 06-18 h: 0.0235 encounters per flight hour
flying only 11-14 h:  0.0077 encounters per flight hour
reduction factor 3.1x

Module lab

Common mistakes

Summary

• The initial ARC comes from the 12 airspace types in SORA Annex C
• Residual ARC sets the TMPR: d High, c Medium, b Low, a no requirement
• The density model compares options and shows that encounters really happen
• Restricting height, route and time is strategic mitigation that needs evidence

Check your understanding

1. Flying below 500 ft in uncontrolled airspace over a rural area gives what initial ARC?
2. What TMPR does residual ARC-c require?
3. An encounter rate of 0.01 per hour over 100 flight hours gives what chance of at least one?
4. If traffic density halves, how does the encounter rate change?
5. Why does BVLOS need additional tactical mitigation?