CASEVAC
UGV

The CASEVAC Problem Under Fire

A soldier is wounded 200 meters from the nearest cover. Enemy machine gun covers the approach. Standard CASEVAC procedure: two soldiers low-crawl to the casualty, apply first aid, drag the casualty back. Time: 10-20 minutes. During those minutes, all three are exposed. Ukrainian data 2023-2024: in contested sectors, CASEVAC attempts produce an average of 0.8 additional casualties per evacuation. In high-intensity fighting, the ratio exceeds 1.0 — each evacuation attempt wounds more soldiers than it saves. At that point, commanders face an impossible choice: leave the casualty or sacrifice more soldiers trying to retrieve them.

The UGV breaks this calculus. It drives to the casualty at 5-10 km/h, takes fire that would wound a human, and returns with the casualty. The UGV has no vital organs. Rifle rounds penetrate the chassis but do not stop the motors. An FPV hit destroys the vehicle (€15,000-50,000) but kills no one. The cost calculation: one UGV lost versus two medics wounded. The UGV is replaceable. The medics are not.

Platform Options

Clearpath Husky (€15,000): 75 kg payload, 4×4 wheeled, 1 m/s on flat terrain. Compact enough for trench systems. Limited by wheel size on rough terrain — gets stuck in mud deeper than 15 cm. Best for paved roads, packed earth, frozen ground. ROS2 Nav2 stack included.

Milrem THeMIS (€150,000): 750 kg payload, tracked, 3 m/s. Military-grade platform designed for CASEVAC. Carries two stretchers simultaneously. Tracked chassis crosses trenches, mud, and snow that wheeled platforms cannot. Integrated NATO STANAG 4609 Ed. 4 (motion imagery) and STANAG 4671 (UAV airworthiness) interfaces. The premium option — justified for units that conduct CASEVAC daily in contested terrain.

Modified civilian ATV (€5,000-8,000): lowest cost option. Convert a commercial ATV (Polaris Sportsman, Can-Am Outlander) with autonomous driving kit (LiDAR + ROS2 + motor controllers). Payload: 150+ kg. Speed: 20 km/h on trails. Not armored but fast — speed is its protection. Requires engineering effort for autonomous conversion.

Lisa 26 Route Planning for CASEVAC

Lisa 26 L2 generates the CASEVAC route using three data layers: known minefields (from brigade S2 intelligence and drone ISR), enemy observation drone positions (from SDR detection and Fischer 26 tracking), and terrain trafficability (from DEM analysis — slope, surface type, obstacles). The route avoids all three threats while minimizing distance. If the shortest path crosses a suspected minefield, Lisa 26 routes around it even if this adds 200 meters and 2 minutes. A destroyed UGV with a casualty on board is worse than a delayed evacuation.

Real-time route updates: if a new threat appears during transit (enemy drone enters the sector), Lisa 26 reroutes the UGV automatically. The UGV stops, recalculates, and continues on the new path. No human intervention needed — the medic at the collection point simply waits longer. If the route becomes completely blocked (threats on all sides), Lisa 26 alerts the company commander for a tactical decision: wait, use suppressive fire to clear a corridor, or accept the risk.

Limitations

Speed: 5-10 km/h means a 500m evacuation takes 3-6 minutes. Human CASEVAC teams cover the same distance in 2-4 minutes under fire (adrenaline). The UGV is slower but survives. In time-critical injuries (arterial bleeding, tension pneumothorax), every minute matters. The UGV does not replace TCCC first aid at the point of injury — a buddy must still apply the tourniquet before the UGV arrives. The UGV replaces only the transport phase.

Terrain: wheeled UGVs fail in deep mud, steep slopes above 30°, dense forest without paths, and water crossings. Tracked UGVs handle these better but cost 10× more. Winter: LiDAR performance degrades in snowfall (snowflakes create noise, reducing effective range from 30m to 10m). GPS-denied navigation adds position drift — the UGV may arrive 20-50m from the intended point and require visual guidance for the final approach.

CASEVAC UGV Platform Comparison

Implementation

# CASEVAC UGV Route Planner — Lisa 26 integration
# Uses ROS2 Nav2 with custom cost layers for threat avoidance

import numpy as np
from typing import List, Tuple

class CASEVACRoutePlanner:
    """Plan UGV route from casualty point to CCP, avoiding threats."""

    def __init__(self, terrain_dem, mine_map, drone_positions):
        self.terrain = terrain_dem        # Digital elevation model (m)
        self.mines = mine_map              # Boolean grid, True = known mine
        self.drones = drone_positions      # List of (x, y, threat_radius_m)

    def cost_at(self, x: float, y: float) -> float:
        """Return traversal cost at grid cell (higher = avoid)."""
        # Hard constraints
        if self.mines[int(x)][int(y)]:
            return float('inf')   # Never cross known mines
        # Slope penalty
        slope = self._compute_slope(x, y)
        if slope > 30:            # Too steep for wheeled UGV
            return float('inf')
        cost = 1.0 + slope / 10.0  # Gentler slopes preferred

        # Drone observation threat — exponential falloff
        for dx, dy, r in self.drones:
            dist = np.hypot(x - dx, y - dy)
            if dist < r:
                cost += 50.0 * np.exp(-dist / r)
        return cost

    def _compute_slope(self, x, y):
        # Sobel-like slope from DEM
        return 5.0  # placeholder

    def plan(self, start, goal) -> List[Tuple[float, float]]:
        """Return waypoint list from start to goal minimizing cost."""
        # Would call Nav2 planner via ROS2 action client
        # Placeholder straight-line return
        return [start, goal]

# Example usage
planner = CASEVACRoutePlanner(
    terrain_dem=np.zeros((1000, 1000)),
    mine_map=np.zeros((1000, 1000), dtype=bool),
    drone_positions=[(450, 320, 150)]  # One enemy drone, 150m threat radius
)
route = planner.plan(start=(100, 100), goal=(800, 800))
print(f"Route waypoints: {len(route)}")

← Part of Platoon Integration

Related Chapters

Sources

Parameter sources. Clearpath Husky specifications (75 kg payload, 1 m/s, €15,000) — manufacturer datasheet clearpathrobotics.com. Milrem THeMIS specifications (750 kg payload, 3 m/s, €150,000) — manufacturer datasheet milrem.com. ROS2 Nav2 parameters — ROS2 Navigation Stack documentation. TCCC guidelines — published by CoTCCC (Committee on Tactical Combat Casualty Care) 2024.

Operational estimates — based on public data, not validated by FSG-A. The "0.8 additional casualties per evacuation in contested sectors, >1.0 in high-intensity fighting" figure is cited from public Ukrainian CASEVAC reports 2023–2024, not independently verified. Modified ATV options at €5,000–8,000 are market estimates, without an FSG-A-confirmed conversion. FPV-strike loss of €15,000–50,000 on a UGV is an estimate based on platform cost ranges. The "20–50 m position drift without GPS" and "10 m effective LiDAR range in snow" figures are engineering estimates, not measured by FSG-A under controlled conditions. FSG-A has no UGV — the module is conceptual and based on manufacturer public specifications.

Economic framing — bounded. The "one UGV versus two wounded medics" comparison simplifies deeply non-linear ethical calculations. It is intended as a rough justification, not as a formula for operational decisions. Commanders always retain authority to weigh the full set of risks in a specific situation rather than mechanically follow this comparison.

External standards and references. Clearpath Robotics Husky specifications. Milrem THeMIS CASEVAC datasheet. ArduPilot Rover. ROS2 Nav2 documentation. Ukrainian CASEVAC casualty data 2023-2025. TCCC guidelines (CoTCCC, 2024). Swedish Armed Forces sjukvårdstjänst.