tesla-radar-driver

Standalone driver for the Tesla Bosch MRRevo14F radar, with the same API as toyota-radar-driver so the same consumer code works with either radar.

Hardware

• Radar: Bosch MRRevo14F (Tesla part, any Model S/X/3 unit)
• CAN adapters:
  • Linux/RPi: Waveshare CAN hat or any socketcan adapter
  • macOS: CANalyst-II USB (canalystii interface)
• Single CAN bus connection (radar CAN1)

Quick start

# Linux (RPi with socketcan)
python3 radar_callbacks.py --channel can0

# macOS (CANalyst-II)
DYLD_LIBRARY_PATH=/opt/homebrew/opt/libusb/lib python3 radar_callbacks.py --channel 0 --interface canalystii

# Curses visualization
python3 radar_curses.py --channel can0

Dependencies

python-can
cantools

For CANalyst-II on macOS: pip install canalystii pyusb and brew install libusb.

API

from tesla_radar_driver import TeslaRadarDriver, TeslaRadarConfig, RadarTrack

config = TeslaRadarConfig(channel="can0")
driver = TeslaRadarDriver(config)

# Register callbacks (same API as ToyotaRadarDriver)
driver.register_track_callback(lambda track: print(track))
driver.start()

# Or poll tracks directly
tracks = driver.get_tracks()  # Dict[int, RadarTrack]

RadarTrack (toyota-radar compatible)

Field	Type	Description
track_id	int	0-31
long_dist	float	Longitudinal distance (m)
lat_dist	float	Lateral distance (m)
rel_speed	float	Relative speed (m/s)
new_track	int	1 if first appearance
timestamp	float	time.time()
raw	dict	All decoded DBC fields

TeslaRadarTrack (extended)

Subclass of RadarTrack with typed access to Tesla-specific fields:

Field	Type	Description
prob_exist	float	Detection probability (0-96.875%)
long_accel	float	Longitudinal acceleration (m/s^2)
lat_speed	float	Lateral speed (m/s)
object_class	int	Classification (0-7)
moving_state	int	0-3
tracked	int	Tracked flag
valid	int	Valid flag
meas	int	Measurement flag
prob_obstacle	float	Obstacle probability
length	float	Object length (m)
dz	float	Height offset (m)

Differences from toyota-radar

Aspect	Toyota	Tesla
Bus count	2 (car + radar)	1 (single bus)
Track count	16	32
Track signals	5 basic	16+ with classification
Keepalive	ACC_CONTROL + 9 static	~30 vehicle state msgs at 100Hz
VIN	Not needed	Auto-read via UDS at startup

Files

File	Description
tesla_radar_driver.py	Main driver (same API as toyota_radar_driver.py)
tesla_radar_protocol.py	Keepalive protocol (~30 CAN message types at 100Hz)
uds_can.py	UDS/ISO-TP for VIN auto-read
radar_callbacks.py	Example: callback logging
radar_curses.py	Example: curses visualization
opendbc/tesla_radar.dbc	Tesla radar DBC
opendbc/tesla_can.dbc	Tesla vehicle CAN DBC (for keepalive encoding)