Behaviour Showcase Scenario

The scenarios/behaviour-showcase.json setup demonstrates independent arm control, local contact sensing, obstacle avoidance and grasping in one scene.

Load and play

From the project directory:

./scripts/run.sh scenarios/behaviour-showcase.json

The scene opens paused. Press Space to play or pause, N to advance one step while paused, and D to restore the original setup, paused.

What to watch

Behaviour	Arm / object	What happens
Multi-arm control	Arms 1, 3, 5 and 7	Four arms move towards separate objectives. Arms 5 and 7 complete unobstructed reaches while Arms 1 and 3 continue their own tasks.
Local sensing	Arm 1 and the food above the body	Contact readings appear when the arm touches the food. Sustained contact by adjacent segments starts the grasping response. Select Arm 1 with 1 to inspect its sensing and status.
Obstacle avoidance	Arm 3 and the circle to the lower right of its attachment	The arm negotiates the obstacle on its curved approach and reaches the target at [532, 460]. Select it with 3 to watch its Avoiding status and progress.
Grasping	Arm 1 and final-meal	Arm 1 establishes a grip, stops and holds the food. The capture count increases. Its slower movement makes this the final interaction.

Arms 2, 4, 6 and 8 stay idle to keep the active behaviours easier to follow. Each active arm has its own target. Arm 1 turns at 2 degrees/second per joint; the other active arms use 40 degrees/second. These are joint speed limits, not whole-arm rotation speeds.

Viewing the individual behaviours

For a general view, play the whole scene and watch the independent reaches. For a closer view of avoidance, restart with D, select Arm 3 with 3, then play. For sensing and grasping, restart, select Arm 1 with 1, and watch its contact indicators and status as it approaches the food. Pause near contact and use N to inspect the transition in small steps.

The sensing and grasping stages are connected: this model senses contact, rather than detecting food at a distance. A grip requires sustained contact from at least two adjacent segments. Contact and Grasping are brief stages before Holding; they are not separate long-running tasks.

For an optional, more visible demonstration that the food is attached, select Arm 1 after it is Holding and press T to retract with the food. Alternatively, click a new destination to carry it there. These are manual follow-up actions, not commands stored in the scenario.

Sequence and verification

The file describes a starting setup rather than a timed script. Independent movement and obstacle avoidance overlap; Arm 1's sensing and capture follow the other completed reaches. To present sensing before avoidance in a video, use separate runs or shots of the same setup in the desired order.

A 15-second simulation check confirmed that:

• All four assigned arms moved, while the four idle arms retained their starting poses.
• Arm 3 entered Avoiding and then reached its target within the normal 2-pixel tolerance.
• Arm 1 progressed through food contact, Grasping and Holding.
• Exactly one capture occurred, after Arm 3 had reached its target.

The scenario does not automatically carry food after capture. Holding is its final state. See Scenarios for a guide to the available setups. Consult the repository’s scenario documentation for accepted fields, units and parameter ranges.