In a world where automobiles are becoming increasingly autonomous, and where semiconductors are a key innovation factor, sensors become crucial, especially when dealing with moving objects. Radar sensors use the FMCW (Frequency Modulated Continuous Wave) technique to reliably detect moving or stationary targets such as cars, trains, trucks and loads in extreme weather conditions. They are also the ideal solution to avoid collisions on mobile machines such as reach stackers, forklift trucks, and port sector machinery such as trolleys, manipulators, and loaders.

Consolidated applications of continuous-wave frequency modulation (FMCW) in the automotive radar sector range from safety to comfort functionalities can be the following: blind spot detection, lane change assistance, monitoring of the driver's vital signs, free space detection, and parking assistance system. These features are based on the radar's ability to detect and locate obstacles in an accurate way, regardless of weather and ambient light conditions.

The high resolution in range and velocity (or Doppler), typical of FMCW radar, makes it suitable for a touchless interface based on gestures. Use cases in the automotive industry include gesture-based door/trunk opening and gesture-based control for infotainment system (such as shaking a hand to switch between screens or rotating a finger to control volume).

The continuous and accurate monitoring of vital driver signs, such as heart rate and respiratory rate is essential for improving road safety. The small dimensions of these sensors imply a non-intrusive implementation; for example, the sensor can be incorporated into the back of the driver's seat.

Continuous-Wave Frequency Modulation (FMCW)

A possible block diagram of an FMCW radar with a single transmission chain (TX) and a single reception chain (RX) is shown in figure 1. The local oscillator (LO) generates a signal, called chirp, with a linearly variable frequency over time and transmitted via the TX antenna. The signal received by the RX antenna (reflected from the scenario in front of the radar) is mixed with the transmission signal to create an intermediate frequency signal (IF). An analog-to-digital converter (ADC) then digitizes the received IF signal for subsequent numerical processing (DSP). The FFT (Fast Fourier Transform) processing, performed on the digitized samples, makes it possible to assess the range of the objects.

Fig 1 Figure 1: block diagram of an FMCW radar (on the left); in the range-FFT plot (on the right), the peak frequencies correspond to the range of the various objects on the scenario [Source: Texas Instruments]

While the frequency of a peak in the FFT plot corresponds directly to the range of the object, the phase of the peak is extremely sensitive to small changes in the position of the object. This sensitivity is the basis of the radar's ability to estimate the vibration frequency of an object and is also the foundation for velocity estimation.

Metrics on the radar performance depend on the choice of the transmission signal. For example:

  • range resolution improves increasing the chirp bandwidth;
  • velocity resolution improves increasing the frame duration;
  • the maximum measurable velocity is inversely proportional to the spacing between adjacent chirp;
  • the number of TX/RX antennas limits the resolution of the angle.

Applications

A free space sensor exploits the radar resolution at large distances and its ability to detect obstacles at close range (poles, walls, cars parked near). A free space sensor can also act as a parking sensor.

The device processes the data coming from the analog-to-digital converter (ADC) along a frame by executing a 2D FFT, which solves objects in range and Doppler, and separates near moving objects from stationary obstacles. With a moving radar, like the one mounted on a door, Doppler resolution also helps to detect objects which, although stationary, are at a different relative speed than the radar. The non-coherent accumulation of 2D FFT arrays creates a Range-Doppler heat map that can be processed by a detection algorithm (figure 2).

Fig 2 Figure 2: typical processing chain for free space sensor applications [Source: Texas Instruments]

The choice of the antenna configuration and of the antenna elements field of view (FOV) are important considerations in free space sensor applications. Generally, a compromise can be found between FOV elevation and ground clutter suppression, as well as between the ability to estimate elevation and the azimuth resolution.

The phase of the signal received in an FMCW radar is extremely sensitive to small changes in the position of the object. By taking advantage of this property, it is possible to estimate the frequency of vibration of objects (such as vibrations induced by heartbeats and breaths). The device transmits a chirp sequence, whereas a peak in the range-FFT identifies a strong reflex coming from the driver's chest. The algorithm in the device tracks the phase of this peak among the chirps and performs a spectral analysis on this sequence of steps to extract the driver's heart rate and breathing rate.

In a gesture-based recognition application, the device performs a 2D FFT on the ADC data collected through the chirps of a frame (figure 3). This solves the scenario in range and in Doppler. Then, the 2D FFT matrix is calculated for each RX antenna (or each virtual antenna if the radar is in MIMO mode). The non-coherent accumulation of the 2D FFT matrix through the antennas creates a Range-Doppler heat map. The next step is to extract more features from the Range-Doppler heat map.

Fig 3 Figure 3: block diagram for a signal processing chain used in gesture recognition applications [Source: Texas Instruments]

We know the damages, even lethal, that can be suffered by children and animals left inside closed vehicles in the heat. An FMCW radar installed in the passenger compartment can detect their presence in an unattended vehicle, thus enabling timely intervention. This application depends mainly on the radar ability to achieve an excellent speed resolution. The radar must separate the objects considering also the minimal movements (like a sleeping child), compared to the stationary clutter in the vehicle.

Sensors ICs

The AWR1x and IWR1x series of sensors are based on CMOS (complementary metal-oxide-semiconductor) technology. The AWR series is dedicated to the automotive sector, while the IWR series is suitable for industrial applications. The mmWave (millimeter-wave frequency band) radars can transmit signals with a wavelength in the order of magnitude of millimeters, and a mmWave system operating at 76-81 GHz, with a corresponding wavelength of about 4 millimeters, has the capacity, to detect movements as small as a fraction of a millimeter.

Each chip enables intelligent and high-precision autonomous detection, with a resolution of less than 4 centimeters, a field accuracy of less than 50 micrometers, and range up to 300 meters. The objectives of the mmWave AWR1x range is to help engineers overcome the obstacles they usually encounter in the design of functions that comply with the regulatory standards of ADAS (advanced driver assistance systems).

77 GHz radars improve automotive safety by allowing vehicles to identify dangerous situations and prevent accidents. They are used to detect different types of obstacles such as other vehicles and pedestrians within a range of 30 to 250 meters, even in poor visibility conditions.

The information provided by radars is used in the ADAS system responsible for multiple applications, including autonomous emergency braking and adaptive cruise speed control. ST Microelectronics has its own STRADA770 transceiver, which covers the millimeter-wave frequency band from 76 to 81 GHz, and includes three transmitters, four receivers, and a twitter modulator.

Fig 4 Figure 4: Radar solution for Automotive [Source: Infineon]

Infineon offers a system solution for 77 GHz radar systems for vehicles that reduces the number of required components (Figure 4). The Radar System IC series (RASIC ™) for the 76–77 GHz range for automotive radars offers a high level of integration. The Infineon 32-bit AURIXTM mutant ADAS device offers a dedicated feature set for radar applications that, in many cases, makes the addition of additional DSPs, SRAMs and external ADCs obsolete.

Conclusion

The diffusion of ADAS systems (Advanced Driver Assistance Systems) - such as electronic stability control, rear-view cameras, and vision-based pedestrian detection systems - has been made possible by the improvements introduced in microcontrollers and sensors. More advanced radar-based embedded solutions provide complementary security features to ADAS system designers.

Capable of supporting a range of new features and functions, 77 GHz automotive radar systems have a promising future. The enormous advantages: high precision and excellent scalability from short to long-range. The downside: a higher degree of technical complexity that is solved with the availability of development kits. The Radar sensors family can be used to implement cost-effective gesture recognition solutions. The sensors of Texas Instruments, ST Microelectronics, and Infineon could be an appropriate choice for single-gesture applications like opening a car door or trunk.