IWR1843BOOST: suggest algorithms for stationary vehicle, human and objects detection(Zero doppler detection) and tracking

Tool/software:

I’m working with the TI IWR1843BOOST mmWave radar without the DCA1000EVM. All processing will be done solely in MATLAB—no CCS, it has to be comprehensive

Application goal

Detect and track vehicles, pedestrians, and miscellaneous objects, including those that become stationary for extended periods.

Scenario parameters

• Object count (n): 1 – 20 vehicles moving on an irregularly shaped roadway.
• Speeds: 30 km/h – 70 km/h before stopping.
• Stationary period: 5 – 15 minutes at zero velocity.
• Detection range: 50 m – 150 m.
• Zones of interest: Rectangular regions, 8 – 12 m long × 8 – 10 m wide for atleast 10-15 zones.

Algorithmic needs

1. Clustering & stationary detection

• Identify and cluster objects.
• Confirm objects that remain motionless for at least 10–15 minutes.

1. Directional tracking

• Track direction and motion history of each object.

1. Zone monitoring

• Define multiple 8 × 10 m (approx.) zones and determine object presence and dwell time inside each zone.

Could you recommend MATLAB-friendly algorithms (or toolboxes) that meet these requirements, or at least serve as a solid starting point for testing and implementation?

Answer 1

[Vinay]

Hi @VLV,

The example "Highway Vehicle Tracking Using Multi-Sensor Data Fusion" demonstrates how to track vehicles on a highway using commonly deployed sensors like radar, camera, and lidar.

This example utilizes the following MATLAB toolboxes and add-ons:

• Sensor Fusion and Tracking Toolbox
• Computer Vision Toolbox
• Radar Toolbox
• Lidar Toolbox
• Computer Vision Toolbox Model for YOLO v4 Object Detection
• Lidar Toolbox Interface for OpenPCDet Library

To run the example, enter the following command in the MATLAB command window:

openExample('fusion/HighwayVehicleTrackingForAutomotiveApplicationsExample')

For additional reference on object tracking, you can also explore:

• https://www.mathworks.com/help/vision/ug/tracking-pedestrians-from-a-moving-car.html
• https://www.mathworks.com/help/vision/ug/motion-based-multiple-object-tracking.html

Hope you find these resources useful!