Adding a New Control Algorithm

Kompass control algorithms live in kompass-core and are dispatched by the Controller component through enum-based registries (ControlClasses / ControlConfigClasses).

There are two ways to add a new algorithm, depending on where the core logic lives:

Python Algorithm

Pure Python or Python + C++ wrapper

Implement your algorithm entirely in Python by inheriting ControllerTemplate, or write a Python wrapper around an existing C++ backend by inheriting FollowerTemplate. Covers configuration, registration, single vs. multi-command (MPC-style) output, and publishing modes.

Adding a Python Control Algorithm

C++ Algorithm

C++ implementation with nanobind bindings

Implement a high-performance path-following algorithm in C++ by extending the Follower class in kompass_cpp. Covers the C++ class hierarchy, parameter system, nanobind bindings, and the build pipeline.

Adding a C++ Control Algorithm

Architecture at a Glance

Layer	Role
kompass	ROS 2 lifecycle node (Controller), topic I/O, config bridging
kompass-core	Algorithm Python classes, config (attrs), registry enums
kompass_cpp	Performance-critical C++ implementations, exposed via nanobind

Existing Algorithms

ControllersID	Algorithm	Base Class	Config Class
DWA	Dynamic Window Approach	FollowerTemplate	DWAConfig
PURE_PURSUIT	Pure Pursuit	FollowerTemplate	PurePursuitConfig
STANLEY	Stanley Controller	FollowerTemplate	StanleyConfig
DVZ	Deformable Virtual Zone	FollowerTemplate	DVZConfig
VISION_IMG	Vision Follower (RGB)	ControllerTemplate	VisionRGBFollowerConfig
VISION_DEPTH	Vision Follower (RGB-D)	ControllerTemplate	VisionRGBDFollowerConfig

Which approach to choose?

• Python-only algorithm (e.g. a proportional controller, a vision-based follower) – use ControllerTemplate. No C++ needed.

• Python wrapper for an existing C++ follower – use FollowerTemplate. Gets path management for free from the C++ backend.

• New high-performance path follower from scratch – implement in C++ first (C++ guide), then wrap in Python (Python guide).