IWR1443: Generic development flow to define/implement the RF and DFE configuration for the specific application

TI’s mmWave sensors are based on Frequency Modulated Continuous Wave (FMCW) radar technology. The design process for an FMCW radar system starts with determining the sensing requirements for the application such as the type of target(s) to be detected, Field of view (azimuth and elevation), maximum range, maximum velocity, range resolution, velocity resolution and angular resolution. The type of target(s) to be detected is quantified as the RCS (Radar Cross Section) value which is a measure of the target’s ability to reflect Radar Signals in the direction of the radar receiver and field of view is dependent on the Antenna design, but the other requirements such as max range, max velocity etc are achieved by designing appropriate chirp and frame design.

The mmWave Sensing Estimator helps users calculate the chirp parameters for TI’s mmWave Radar devices based on high level inputs such as maximum range, range resolution, maximum unambiguous velocity etc. (If you need to understand the theoretical aspects behind the various chirp parameters and their relation to the scene requirements, please look at the app note Programming Chirp Parameters in TI Radar Devices)

Next, you can enter the various parameters calculated by the Sensing estimator into MMWAVE-STUDIO to evaluate the detection performance of the chirp configurations using MMWAVE-STUDIO. Please refer to the MMWAVE-STUDIO users guide for more details. You can also evaluate the sensor using pre-defined configurations with the mmWave SDK OUt of box demo and mmWaveDemoVisualizer.

The computed chirp parameters are programmed into the mmWave device using the mmWave Link API which is provided in TI's mmWave SDK. The mmWave Link API allows the application to configure, control, and monitor the Radar subsystem using high level function calls. Hence, the next step is to map the chirp parameters to a set of mmWave Link API calls which are called from the application. To understand this process, please refer to the mmWave SDK users guide, especially the section "Configuration File format". Most of the lines in the Configuration file map to an equivalent API that configures the corresponding parameters in the Sensor. The CLI in the demo parses the Cfg file and maps the parameters to the API calls (for e.g. profileCfg., chirpCfg, frameCfg, etc).

With the sensor configured according to the scene requirements, the next step is to design the digital processing chain to process the ADC data coming from the front-end to extract the range, velocity and angle information and optionally perform other high level operations. The typical FMCW processing flow on the received data consists of a series of FFT operations for Range and Doppler processing, followed by detection algorithms (such as CFAR) and Angle of Arrival estimation. TI’s IWR1443 device includes a Hardware Accelerator for FFT (Range, Doppler and Angle estimation) and CFAR detection processing. The IWR1643 replaces the Hardware Accelerator with a C674x DSP which is used for the FFT (Range, Doppler and Angle estimation) and CFAR detection and also to run advanced algorithms such as clustering and tracking. The mmWave SDK provides the hardware accelerator drivers for the 1443 as well as software FFT and CFAR libraries for the DSP on 1642 along-with other drivers such as ADC buffer, SPI, HWA, and EDMA etc. to help implement the radar processing chain on either device. Please refer to the SDK OOB demos and the various other examples provided under mmWave Industrial Toolbox to understand the software processing.

Please also refer to mmWave Training Series for many helpful training resources.

Regards

-Nitin