Cooperative path following · 2–5 vehicles
Compare independent line-of-sight guidance, independent virtual targets, and coordinated virtual targets. Give each robot a different path, configure wingmen or leader–followers, and disturb one vehicle to reveal why path following alone is not formation keeping.
1 · Experiment setup
Path edits restart · Speed adjustments apply live2 · Vehicle paths and formation slots
Path and formation parameters are independent.Sine x-shift: moves the waveform horizontally (positive values shift peaks right). It is not the same as a wingman’s desired lag. VT pace ×: multiplies the global target speed for one robot and can be adjusted during a run; formation coordination may override the resulting free-running pace.
3 · Live top-down simulation
t = 0.0 s / 110 sError histories
Vehicle x-station (green) · target x-station (amber) · vehicle–target distance (blue)Instructor's 3-minute demonstration
1. Turn speed-limit fault OFF. Use coordinated VT with “Track actual virtual target.” Watch blue vehicle-to-target error converge near zero while the formation metrics do their own work.
2. Choose the “Keep lookahead gap” objective: the target retains an approximately 9.5 m x-station lead; the blue Euclidean distance also depends on path shape. Reset, restore target tracking, then enable the 18 s speed-limit fault to demonstrate residual error under constraints.
3. While running, adjust VT speed U₀. When it exceeds achievable vehicle speed, tracking error increases. Change just one vehicle’s VT pace factor; compare independent and coordinated VT. Then try the shifted sine preset.
Concepts to discuss
Wingman: a slot specifies both its desired lane (lateral offset) and its desired path-station delay behind the leader. Select “Matched wingmen” for a formation on similarly shaped paths.
Leader–follower: select “Leader–followers.” Each robot aims for a different delay on a common route. This is not collision avoidance: the separation metric uses vehicle points and contains no hull geometry.
Different paths: assign individual straight, sine, or curved routes. Change each sine wave’s x-shift and notice how shifting a path differs from requesting an along-path formation lag. Progress synchronizes x-stations—not Euclidean separations. Export CSV for analysis.