Interactive command workbench

Drive. Scan. Decode.

Watch the robot move, follow its LiDAR scan, and inspect every wheel command.

MANUALSIMULATION ONLY

Start here · browser simulation

See FieldScout move in one click.

No real robot is connected. This page moves a simulated robot and displays example messages; it sends nothing to hardware.

Start: press the green button to load a route and begin at 0.50 m/s.
Watch: the view follows the robot at a closer metric zoom. Use Fit arena to see all goals and routes.
Stop: click Stop simulation at any time, or press Space outside text inputs. For your own route, click clear ground and choose Run my waypoints.

Ready. Start example mission loads the route and starts this simulation for you.

MISSION PROGRESSNo route

CURRENT GOALStart an example

DISTANCE TO GOAL—

SIMULATION TIME0.0 s

Metric robot & footprintLaser returns / current routeDashed original routeNumbered waypointsObstaclesClick clear ground to add a waypoint

Position X / Y1.00 / 1.00

Heading0.0°

Linear m/s0.000

Yaw rad/s0.000

Command seq.00000

STOPPED · Start an example mission or choose Start manual driving.

Manual drive

Manual sequence: Click Start manual driving, then hold an arrow or W/A/S/D. Release to stop. A waypoint mission starts when you press its Start button.

Speed setpoint 0.25 m/s

Hold WASD or arrows
Space stop · X emergency stop · release to stop

Waypoint mission

Arena

No custom route yet. Start an example above, or click clear ground to add your own goals.

This button starts your route in the simulation. If a route or fault needs attention, the message above the arena tells you what to do.

Enter a waypoint by coordinates

X Y

Try obstacle avoidance

Start an example mission. Add a box ahead while the robot moves: its simulated LiDAR detects the box, then the planner chooses a detour toward the same goal.

Start a mission to try obstacle detection and detours.

This test restarts in the open arena and adds a full-width barrier after 2 simulation seconds. With every passage blocked, the robot stops; remove the barrier and resume.

If no safe route exists, motion stops. Remove the added obstacles and choose Resume mission. New boxes are shown in amber.

Fault injection & reset

Interrupt command streamRemove LiDAR from autonomy

Command loss exercises a simulated 400 ms watchdog. Clearing a fault does not authorise motion; explicitly start manual driving or a mission again.

01 / Navigation → wheel command

20 Hz target

Body frame: +X forward, +yaw counter-clockwise. Wheel order: FL, FR, RL, RR. Positive logical wheel speed always means vehicle-forward.

v = 0.000 m/s     ω = 0.000 rad/s

CMD,0.000,0.000,0.000,0.000

Newline-terminated ASCII expected by the Lua design. A matching ESP32 receiver is absent from the supplied sketches.

Logical wheel	rad/s	RPM	Driver channel	Raw RPM

Local message downloads are simulation examples. No hardware is connected.

02 / ESP32 → Modbus RTU

115200 · 8N1

Function 0x10 · start register 0x2088 · two signed 16-bit RPM values · CRC low byte first. Wire examples, not transmitted frames.

DRIVER 1 · FRONT LEFT / FRONT RIGHT

DRIVER 2 · REAR RIGHT / REAR LEFT

Expected ACK frames & byte legend

ID 10 20 88 00 02 04 [L hi lo] [R hi lo] [CRC lo hi]. A driver write acknowledgement confirms register acceptance, not wheel motion or odometry. This simulator does not receive hardware acknowledgements.

Recent command messages

Model assumptions

A learning model, openly defined.

Differential-drive kinematics use r = 0.0535 m and track = 0.250 m from the Lua scene, with a 0.30 m circular collision radius. The full drawn chassis and wheels occupy 0.46 × 0.34 m and remain inside this circle. Map dimensions are metric; the separate robot-detail inset is clearly labeled as a close-up. The model omits slip, inertia, real braking and sensor uncertainty. The isometric robot is schematic; the original GLB assembly is in 3D robot.

Autonomy method

Watch a command pipeline work.

A* starts from the known map. A synthetic LiDAR ray that hits an added box reveals that entire box to the planner. The robot stops, plans a detour and continues toward the same goal when a safe route exists. Amber dashes show the initial route; cyan shows the updated route. This is a teaching simplification, not inference of obstacle shape from real point clouds. A waypoint follower uses exact simulated pose and a protective stopping gate. It demonstrates message flow; it does not run SLAM, Nav2 or a real localisation estimator.

Next integration task

Replace each ideal assumption.

Students implement real scan acquisition, odometry, TF, mapping/localisation, planning and the ESP32 command receiver. Validate physical stopping and independent emergency-stop behaviour before powered floor tests. Follow the build plan →