跳至正文

Integration

Drop in a few parts but don't drop your plans.

Adding AWEARE doesn't take a rip-up-and-replace action plan. Minimal hardware changes open the door to revolutionary performance. The waveform generation and DSP back-end software get supercharged.

01 · Architecture

Where the AWEARE keys sit in your existing radar.

Two keys sit on either side of the existing RF chain. One encodes the AWEARE waveform on transmit; one decodes it on receive. Everything else is unchanged.

Radar Control Target Analyzer Tracker Waveform Generator Transmit Chains Signal Processor Receive Chains HW KEY SW KEY AWEARE HARDWARE + SOFTWARE KEYS Same 24 GHz front-end. AWEARE waveform encodes a code that the signal processor uses to discriminate real returns from ghosts.

02 · Integration paths

Three paths, depending on where you are in the stack.

Path A

Companion module

Tier 1s + OEMs with existing 24 GHz radar

  • →Small companion module sits in the transmit chain
  • →AWEARE DSP runs on existing radar processor (or as licensed IP)
  • →Fastest pilot path — weeks, not quarters

2-3× resolution, ~80% fewer ghosts

Path B

Embedded DSP IP

Chipset companies (ST · NXP · TI · Infineon)

  • →AWEARE waveform IP licensed into your DSP back-end
  • →RF front-end remains your design; AWEARE Tx encoder added
  • →First chipset shipping with AWEARE-ready end-to-end signal chain

Embedded license · royalty model

Path C

Next-gen 77 GHz

Imaging-radar product designs

  • →AWEARE 2027 next-gen waveform (3-4× resolution, 90%+ ghost reduction)
  • →Targets ASIC integration in 77 GHz silicon
  • →Co-development with the AWEARE engineering team

Co-development · 2027 milestone

03 · Sensor fusion

Plays nicely with LiDAR and Cameras.

AWEARE pairs naturally with LiDAR (e.g. Mobileye-class systems) and with Cameras fusion. Cleaner radar at the signal layer means downstream perception stacks can use rules-based DSP instead of compute-intensive scene-building AI.

  • → Cleaner radar reduces or eliminates the need for radar-side scene reconstruction.
  • → Pairs with LiDAR for redundant range, with Cameras for redundant classification.
  • → Reference: Liu et al., NeurIPS 2023, Echoes Beyond Points.
RADAR AWEARE 24 GHz LiDAR e.g. Mobileye CAMERA Vision stack FUSION LAYER Rules-based DSP (specular reflection calculators, multipath ghost filters) + much lighter AI scene-building PERCEPTION OUTPUT Clean object list + velocities, ranges, classes For: ADAS · AV · fleet safety policies AWEARE radar pre-filters multipath ghosts at the signal layer, so downstream fusion needs less scene-building AI compute.

04 · Pilot scope

What an AWEARE pilot looks like.

We move in pilots, not quarters. The smallest meaningful engagement is a measured side-by-side test on your hardware in your environment.

Duration

6–12 weeks for a side-by-side; 3–6 months for a chipset embedding.

What you provide

Test radar (or chipset platform), test environment, perception team contact.

What we provide

AWEARE companion module or DSP IP, on-site engineering, measurement protocol.

Outcome

A measurement-grade comparison report you can take to your safety / certification teams.

Engineering-to-engineering

Talk to the engineering team.

The fastest way to evaluate AWEARE is to look at our test data and start a pilot. We are happy to NDA, share measurement protocols and walk your engineers through a side-by-side.

联系工程团队 查看技术细节