camsense-X1

unofficial reverse engineering of a Chinese LiDAR.

Discussed this on my discord: https://discord.gg/zRGJcqa

ROS1 driver for the camsense-X1: https://github.com/Vidicon/camsense_driver Made by Vidicon (Bram Fenijn)
ROS2 driver for the camsense-X1: https://github.com/rossihwang/ros2_camsense_x1 Made by rossihwang

implementation for a STM32: https://github.com/anhui1995/Camsense_X1

Read the LiDAR with a M5Stack: https://github.com/yishii/LiDAR_Camsense_X1_M5Stack Made by yishii

Your opensource camsense-X1 project here? Please let me know.

Dimensions

The dimensions(mm) of the Camsense-x1 mounting holes:

I have made a 3d model of the LiDAR: stl, step (Dimensions are approximate)

Electrical Connections

There are 3 pins needed to connect to the lidar: 5V, GND and TX. For this you can use the mounted connector(part number unknown) or solder some pinheader to the pcb.

The tx pin sends serial data with 115200 baud at a signal level of 3.3v This make the lidar safe to be connected directly to a 3.3v microcontroller of SBC like the STM32 or Raspberry Pi

A FTDI adapter can be used to connect the LiDAR to a computer. The TX pin of the LiDAR needs to be connected to the RX pin of th FTDI adapter. The LiDAR can be powered using the 5V of the FTDI adapter.

Communication Protocol

The lidar sends on average 50 packages per rotation of the sensor. A package is always 36 bytes and has the following format:

<0x55><0xAA><0x03><0x08>
<speedL><speedH>
<startAngleL><startAngleH>
<distance0L><distance0H><quality0>
<distance1L><distance1H><quality1>
<distance2L><distance2H><quality2>
<distance3L><distance3H><quality3>
<distance4L><distance4H><quality4>
<distance5L><distance5H><quality5>
<distance6L><distance6H><quality6>
<distance7L><distance7H><quality7>
<endAngleL><endAngleH>
<unknown><unknown> could be a CRC

A package always starts with <0x55><0xAA><0x03><0x08>

Calculate rotation speed in hz:

float Hz = ((uint16_t) (speedH << 8) | speedL) / 3840.0; // 3840.0 = (64 * 60)

Calculate start and end Angle in degrees:

 float startAngle = (startAngleH << 8 | startAngleL) / 64.0 - 640.0;
 float endAngle   = (endAngleH   << 8 | endAngleL)   / 64.0 - 640.0;

Best method we found to calculate index and parse data to array:

float offset_ = 16.0; // 0 degrees seems to be 16 degrees of center.
const float IndexMultiplier = 400 / 360.0;

float step = 0.0;
if(endAngle > startAngle)
{
    step = (endAngle - startAngle) / 8; 
}
else
{
    step = (endAngle - (startAngle - 360)) / 8; 
}

for(int i = 0; i < 8; i++) // for each of the 8 samples
{
    float sampleAngle = (startAngle + step * i) + (offset_ + 180);
    float sampleIndexFloat = sampleAngle * IndexMultiplier; // map 0-360 to 0-400
    int sampleIndex = round(sampleIndexFloat); // round to closest value.
    index = sampleIndex % 400; // limit sampleIndex between 0 and 399 to prevent segmentation fault
    
    uint8_t distanceL = data[8+(i*3)];
    uint8_t distanceH = data[9+(i*3)];
    uint8_t quality = data[10+(i*3)];
    
    
    if(quality == 0) // invalid data
    {
      distanceArray[index] = 0;
      qualityArray[index] = 0;
    }
    else
    {
      distanceArray[index] = ((uint16_t) distanceH << 8) | distanceL;
      qualityArray[index] = quality;
    }
}