Integer control callbacks lose fractional vehicle limits and stop motion

Aperta
  • Qualcosa non va?
#936 1 commento 0 reazioni 0 assegnatari Vedi su GitHub

I maintainer di solito rispondono entro 1 giorno

@Elon-Jiang ci sta già lavorando.

Dal 10/11/2026.

  • #937 di @Elon-Jiang — aperta

Valutazione

Difficoltà
2/5
Tempo stimato
1-3 ore
Idoneità per principianti
20/100
Tipo di issue
Bug
Chiarezza
Specificata chiaramente
Stato di attività
Ferma
Stack tecnologico
python
Ambito
backend

Direzione di ricerca

Read the three u_limited methods in Vehicle.py and trace how integer control inputs determine the result dtype. Reproduce the reported difference between integer and floating callback commands with the public step example, then add regression coverage for the limit methods and callback path. The issue reports an open linked pull request (#937), so the work is already under way.

Scritto dal modello di indicizzazione a partire dal testo della issue.

Descrizione

RTB version

1.5.0, current main 879fb363580840e19ee70b6b175e8cd3543e2b38 (normal editable install with compiled extensions).

Python version and OS

Python 3.12.13, macOS arm64.

Expected behaviour

An integer-valued control callback should produce the same limited motion as the equivalent floating-valued callback. The documented ArrayLike inputs include integer lists/arrays.

Actual behaviour

u_limited in Bicycle, Unicycle and DiffSteer allocates an integer result for integer inputs, then truncates fractional speed, acceleration or steering limits on assignment. For example, a desired speed of 1 limited to 0.5 becomes 0. This also reaches the public simulator through a control callback:

import roboticstoolbox as rtb

for command in ([1, 0], [1.0, 0.0]):
    vehicle = rtb.Bicycle(speed_max=0.5, dt=0.1)
    vehicle.step(lambda vehicle, t, x: command, animate=False)
    print(command, vehicle.x)

Current main prints [0., 0., 0.] for the integer callback and [0.05, 0., 0.] for the floating callback. Unicycle and DiffSteer reproduce the same lost motion. Direct constant step([1, 0]) is a passing control, since eval_control already converts that route to floating point. No rendering or robot hardware is involved.

I checked all 24 open PR file lists, full history for Vehicle.py, and complete searches for u_limited, limits_va, integer velocity and acceleration. #929 changes the sign of reverse odometry; it does not change control limiting.

I plan a focused floating-output correction in these three limit methods and regression coverage through both u_limited and the public step callback path. Prepared with assistance from OpenAI Codex (GPT-6).

Lingua principale
C++
Stelle
3.6k
Fork
650
Merge medio
14h 6m
PR unite (30g)
139

Preparare l'ambiente

Come iniziare

  1. Leggi tutta la issue e poi la guida ai contributi del progetto.
  2. Commenta sulla issue per dire che te ne occupi tu — evita che due persone facciano lo stesso lavoro.
  3. Fai un fork del repository e lavora su un branch.
  4. Apri una pull request che faccia riferimento al numero della issue.

Altre issue di petercorke/robotics-toolbox-python

Tutte le issue di petercorke/robotics-toolbox-python

Issue simili

Altre issue su C++

Ricevi le nuove issue nella tua casella

Un breve riepilogo di issue GitHub adatte ai principianti.