Using the LEDDAR One on Phoebe, I had to drop her sampling rate to 500Hz to prevent FIFO overflow. This isn’t bad, but I decided to investigate why adding LEDDAR into the mix has slowed the processing down significantly. Out with my OSCIUM iMSO-104 iPad scope.
The following allows a crude estimate of the the baud rate to be at least 50kbps:
This shows the time the LEDDAR One takes for a request / response processing – about 3ms
This shows the reaction time to the rising edge data ready interrupt (DRI); since the Quadcopter and the test code only poll the interrupt every 10ms, this is fine too.
What’s also visible is the cost of doing 3 individual reads of the registers (about 20ms) compared to reading 5 registers in a single shot which drops it down to about 5ms as shown below:
I’ve swapped Phoebe to use the one-shot read as above, and as a result, she’s running the IMU at 1kHz again while running the LEDDAR.