mirror of
https://github.com/pollen-robotics/microduck.git
synced 2026-11-22 08:14:21 +08:00
The brain gains two states, both driven by robotd and wanted by the brain: - Laser: tracking is on whenever the duck could chase (not tired, not napping, petted or playing ball). A followed dot interrupts any standing or walking state; the chase ends 2 s after the dot goes, and costs twice a walk's energy. Below 0.2 the laser goes off and the duck naps; it comes back above 0.45. - BallPlay: picked from the standing pool when rested and off cooldown, less often the more bored of the ball. No ball in 20 s backs off 3 min, doubling per miss up to 30. A game ends on 3-8 kicks, 1-3 min shrunk by boredom, the ball lost for 15 s, or tiredness; then 4-8 min of rest. While robotd drives either, autod sends a zero twist and a centred head, which robotd reads as nobody asking. A game somebody else started is stood aside for, and the brain's own is switched off when the pad takes over. The on/off switch moves into robotd (robot.autonomous, read back in robot.mode and robot.state; part of v59), so the NFC autonomous_mode tag can switch it — nfcd cannot write robotd.toml. [autonomous] enabled is now only where each boot starts; robotctl autonomous on|off switches it now. autod stands the robot up on the switch's rising edge, and once per boot for the switch at boot. padd learns it from robot.mode. mapd starts mapping the first time it sees the duck switched on. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>