• State Resolved
  • Locked Locked
  • Replies 6 replies
  • Subscribers 30 subscribers
  • Views 1274 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

IWR1443BOOST: Is this the right board for my project?

John Kasha
Prodigy 40 points
Part Number: IWR1443BOOST
Other Parts Discussed in Thread: IWR1642BOOST, AWR1642BOOST-ODS, AWR1243

I have a patent on using a sensor on a golf club head to measure the distance, angle, and speed to a golf ball as the golf club head is moving within about 1-2 feet of the golf ball.

First of all, is the IWR1443BOOST the right board to use in testing (as opposed to the IWR1642BOOST)?

Can I get distance, angle, and speed measurements within milliseconds?  I know the radar is fast enough, but can the reading electronics provide multiple readings as the club moves at as much as 100 mph in the 1-2 feet in front of the ball?

I know you have an example of statically measuring a quarter at 2 meters.  I essentially, want to dynamically measure the distance, angle, and speed of a club head relative to a golf ball to get an accurate map of the club head before it strikes the ball.  How accurate will the measurements of distance, angle, and speed be at those speeds?  What is the minimum distance at which I can still measure the distance to the ball (1, 2, 3 inches)?  Will the angle measurement be accurate enough to give me a good picture of the club face relative to the ball?

Thanks.  Any help with this application is greatly appreciated.

John 

  • 0
    TI__Mastermind 25835 points
    Hello,

    This is a nice interesting project, TI's mmWave Radar can measure distance, velocity and Angle information.
    Both IWR1443BOOST, IWR1642BOOST could be used, However these boards have Antenna designed for longer distance 50m, 100m + distance applications. 1443 Device has FFT hardware accelerator and programmable ARM processor, 1642 has programmable ARM processor and 6748 DSP and extra memory for the signal processing. For this application you may need to design custom antenna board with shorter range but higher field of view.

    You may explore www.ti.com/.../awr1642boost-ods board. These antennas provide higher field of view and better angular resolution in both Azimuth and Elevation direction.

    There is a lab in TI TIREX folder which provides additional details on obstacle detection features.
    dev.ti.com/.../

    Measurement resolution would depend on Radar Chirp configurations. You may want to modify the Radar chirping configuration depending upon your target application requirements probably some trade-off would be needed on the chirping configuration, meeting all the requirement would be challenging. A generic chirp configuration details are provided in the below link, you could optimize these chirp configuration depending upon your range resolution, velocity resolution, refresh rate requirements.

    dev.ti.com/.../.

    One of the challenge in this application, I can think of is the Radar cross section of the golf ball, it may be difficult to exactly isolate such small object w.r.t to external clutter environment.

    Radar can measure 6-7 cm distance from the target and also capable of measuring 100 mph as these sensors are being used in Automotive Radar applications as well.

    Another challenge is to meet the distance, angle, and speed measurements within milliseconds. Processing time needed for the DSP to compute distance, angle and speed measurement itself will take few milliseconds. Hence this may not be possible.

    As you could note from the below chirp configuration, Depending upon the application requirements refresh rate can vary 5-50Hz.
    dev.ti.com/.../.

    Thanks and regards,
    CHETHAN KUMAR Y.B.
  • 0 in reply to CHETHAN KUMAR Y.B.
    Prodigy 40 points

    Thanks for the great information.  

    The awr1642boost-ods board is a good suggestion for testing, and I think developing a different antenna would not be a big problem.

    Two comments you made, however, make me think the ti MMwave might not be the right part for me.  I need to get at least 4 measurements within 7 ms. 

    You wrote that "[p]rocessing time needed for the DSP to compute distance, angle and speed measurement itself will take few milliseconds. Hence this may not be possible." 

    Do you mean with the DSP and processor on your evaluation boards, or do you mean not possible ever?

    I understand that signal processing for unknown objects can be time consuming.  However, in my case the ball will always be in about the same place, so shouldn't the processing be much easier?

    In any event, if you do not think this is the right product for me, please let me know.

    Thanks,

    John

  • 0 in reply to John Kasha
    TI__Mastermind 25835 points
    Hello,

    Yes, you could start your system evaluation with the awr1642boost-ods.

    Typical frame rates are 30-50 frames per second. 4 measurement for every 7msec is challenging , this could be limiting factor, within our chip we have C6748 DSP integrated and it would be operating max at 600Mhz. We cannot increase it's speed/MIPS.

    For range/Velocity/Angle estimation we need to perform Range FFT /Doppler FFT/Angle FFT on the received signal after digitization to perform this task for the DSP would take 2-3msec. Probably this could be improved if we use faster external DSP.

    If this is very hard requirement then you may want to review AWR1243 + external DSP. AWR1243 is mmWave front end sensor only and it provides raw ADC data over high speed CSI2 interface which could be connected to external DSP. You may choose faster external DSPs to meet your throughput requirements.

    Golf ball is very tiny object, it's radar cross section would lesser, Hence it would be reflecting with weak energy level (Apart from that, Ball will scatter electromagnet signals in all direction instead of reflecting back to radar, hence received signal would be weaker it would appear as small point object), Isolating this weak target w.r.t to strong other targets, such as reflections from neighboring places/ground reflection would be challenging. If by some means if you could make golf ball make reflect stronger signals towards Radar from which we could isolate uniquely then it's a viable solution.

    Thanks and regards,
    CHETHAN KUMAR Y.B.
  • 0 in reply to CHETHAN KUMAR Y.B.
    Prodigy 40 points

    Thanks again for the quick response.  I am not so worried about the locating the ball.  The ball, tee, and even a mat can be under our control.  We could potentially paint the ball and tee to reflect, and the mat to absorb.  Do you have a recommendation for a type of paint to reflect your signal and a type of mat material to absorb your signal?

    I am still concerned about the timing and thank you for the suggestions.

    One thought I have is that for one application, I do not need to process the information within the 7 ms.  In other words, is there enough memory to store or buffer measurements and then process the measurements later?

    The other thought I have for the real-time application is whether or not the DSP is necessary. IWR1443BOOST does not have a DSP, but it can determine range.  How is it working?  In other words, do I absolutely need the DSP?

    Thanks again,

    John

     

  • 0 in reply to John Kasha
    TI__Mastermind 25835 points
    Hello,

    If you have control over the ball, tee and mat then it could simplify the problem a bit. For the RADAR metals are the good reflectors if you could have metallic ball or outside metallic coating then it would reflect better.

    For the absorbers there are various commercially available materials with different range of frequency, absorption levels and cost variants.
    product.tdk.com/.../e9e_bdj_003.pdf
    media.digikey.com/.../EMI_MicroAbsorb_Cat.pdf
    www.eccosorb.com/products-eccosorb-hr.htm

    If you do not need to process that data within 7msec then it could simply the problem, as you rightly pointed out you could store it in external storage media. Within our devices we have very less memory space less than 1.5Mega bytes. For real time application you could use external DSP with large DDR memory space for processing.

    Alternatively, if you want to do offline processing on a PC framework, we have DCA1000 hardware which could be interfaced to 1443boot/1642boost boards for offline processing.

    You could order this board from following link:
    www.ti.com/.../DCA1000EVM

    This has 2Gbytes of storage space and could stream it to PC over Gigabit Ethernet interface.
    And you could use mmWave studio for the control/post processing.

    In 1443 device we have hardware accelerator which does FFT computation for the Range, Doppler and Angle processing, these are very basic Radar signal processing. if you need to do beyond this such as object tracking classification you would need higher layers of processing requirements which may not be possible with 1443 kind of device. Then DSP may be needed as in 1642 kind of silicon. If you need faster processing/MIPS beyond 600Mhz and more than 1.5 Megabyte memory requirement then you may use the mmWave front end device only to stream the real time ADC data and use faster external DSP and more memory with the DDR interface.

    Thanks and regards,
    CHETHAN KUMAR Y.B.