2024 R1: Ansys AVxcelerate Sensors Adds Beamforming and Arbitrary Waveform Simulation Capabilities

One of the biggest trends (and challenges) in autonomous vehicle (AV) tech for 2024 will be the development and advancement of advanced sensor technology that can drive greater vehicle perception. Autonomous vehicles rely on a range of sensors, including radar, to enable those crucial tasks that automate our driving experiences to help us stay in our lane or avoid obstacles in our path, for example.

Ansys AVxcelerate Sensors simulation software is introducing important new features with its first release in 2024 to further optimize sensors that advance AV perception. This release extends the capability for radar sensor makers, tier ones, and original equipment manufacturers (OEMs) to go one step further to represent the radar model used in simulation. Two important radar waveform propagation techniques (beam forming and custom or arbitrary waveform) can now be used.

Why is this new capability important? Beam forming / steering and arbitrary waveforms are techniques used by radar manufacturers to optimize radar signal processing and to address on-road perception challenges, starting earlier in the design cycle using advanced simulation. It is one of several new features that support the main pillars of AVxcelerate Sensors, a solution dedicated to the development, testing, and validation of advanced driver assistance systems (ADAS) and autonomous driving (AD) technology with safety in mind.

AVxcelerate's Beamforming Simulations Steer Signals in a Desired Direction

Beamforming radar systems use antenna arrays to focus or steer radio signals in a specific direction, without any mechanical actuator. Beamforming, or the act of combining signals in the desired direction, improves radar signal quality, range, and resolution.

In the latest release of 2024 R1 AVxcelerate Sensors, you can feed a radar simulation with a complex weighting sequence for each antenna or each of the system chirps or pulses to optimize radar beamforming capabilities. This sequence can then be modified in AVxcelerate Sensors in real time per digital signal processing (DSP) software request.

Following a simple sequence of steps, you can first design the characteristics of the resultant radiation field. Next, you can create the associated transmitter weightings sequence for each radar configuration. Then, in AVxcelerate Sensors, it's possible to create or switch one mode per configuration or update the transmitter weightings sequence at each step during the simulation.

Virtually adjusting beam patterns in AVxcelerate Sensors is particularly useful for enhancing the performance of radar-based blind spot warning systems, a typical challenging scenario for radar sensor performance.

To illustrate, consider a motorcycle riding in the blind spot of an AV moving at the same speed. In this scenario, the AV is equipped with a long-range radar that is unable to pick up the faint echo of the motorcycle, which is mixed up with ghost signals coming from guardrails and other objects in the periphery. Due to a system weakness point, the AV will fail to detect the motorcycle, creating a potentially dangerous situation.

Simulating radar beam forming based on the same radar hardware in AVxcelerate Sensors, you can virtually adjust the beam pattern strategy by inputting different weightings to the antennas. This enables you to alternately switch between long range and blind spot-optimized detection radar modes to pick up a stronger reflection from the motorcycle and filter out any ghost signals coming from other sources. The ability to simulate various configurations of the radar enables developers to optimize their software algorithms early in the development cycle, saving valuable time and cost during the design process.

All of this is made possible by AVxcelerate Sensors' beamforming capabilities, which enable you to input a complex weighting sequence for each antenna chirp or pulse in order to define the passive intermodulation (PIM) shape around the emitter. During this simulation, weighting sequences can also be modified at each simulation step to create a change in the beam shape or to create a virtual scanning motion that improves system perception.

Leveraging the combination of Ansys HFSS 3D high-frequency structure simulation software and AVxcelerate Sensors, the creation of an initial and subsequent complex weighting sequence for antennas can be efficiently generated in HFSS, taking into account interactions generated by the mounting on the sensor on the vehicle.

AVxcerelate's Arbitrary Waveform Capabilities Mitigate Signal Jamming Challenges

Also new for 2024 R1 AVxcelerate Sensors are its radar arbitrary waveform simulation capabilities. Radar arbitrary waveform is a technique used by radar sensor manufacturers to solve signal jamming challenges caused by several radars operating in the same area. Jamming often happens on the same road between vehicle radars operating on the same frequency band that cross each other.

Other radar interferences can also jam the signal between the emitted and received signal and create what's called false targets, which lead to processing errors that can result in unplanned collisions or unexpected breaking. To avoid jamming during radar detection, radars often use random frequency hopping and random timing change methods to “sign” the waveforms so the jamming signal can be filtered out before processing the range-doppler maps and angle of arrival.

To this end, radar makers “sign” or mark their waveforms with a unique sequence of frequencies to filter them out of the ambient noise and perturbations at the reception stage in a process known as interference mitigation. In the 2024 R1 AVxcelerate Sensors release, users can now simulate radar sensors emitting those unique sequences in both arbitrary frequency modulated continuous wave (FMCW) or arbitrary pulse-doppler waveforms. This encoding and decoding processing is the core of proprietary digital signal processing.

AVxcelerate Sensors’ arbitrary waveform capabilities for FMCW radars and pulse-doppler waveforms enable you to base a very simple extraction of a jam signal that can then be recognized to extract only an emitter signal. This enables radar manufacturers to use their final DSP that considers arbitrary waveform filtering in simulation. For simulation accuracy and credibility, it is important that the signal can be encoded and decoded to properly test the real digital signal processing algorithm as part of the simulation loop.

Of course, AVxcelerate Sensors’ capabilities don't stop here. Whether you're a current user or simply curious how we can help you reach your AV development objectives, visit the AVxcelerate Sensors product page for more information.