Collecting measurements from multiple sensors wirelessly and simultaneously

Sensors Distance
1

I want to obtain depth measurements from multiple (~30–50) dToF imagers (e.g., SparkFun Qwiic Mini dToF Imager - TMF8821). In addition to depth measurements, I also want to determine the orientation of each dToF imager using an accurate IMU sensor (e.g., SparkFun 9DoF IMU Breakout - ISM330DHCX, MMC5983MA) attached to each imager. My goal is to collect both orientation and depth measurements from multiple sensors simultaneously.

I believe I will need multiple microcontrollers to handle these measurements. Using one microcontroller per sensor may be inefficient. Additionally, since these sensors will be placed on moving objects (e.g., robot vacuum cleaners), a wireless connection will be necessary. Therefore, the microcontrollers should be compact and support Wi-Fi or Bluetooth. I may also need multiple batteries for wireless connectivity.

Here are my questions:

  1. What components are necessary for this project?
  2. Could you recommend compact, wireless microcontrollers (perhaps the SparkFun Thing Plus - ESP32 WROOM (USB-C)?) suitable for this project?
  3. How many sensors can be connected to each microcontroller while ensuring simultaneous depth and orientation measurements?
2

Please explain what you mean by taking measurements simultaneously.

A processor reads one sensor at a time and it takes time to store or send the data elsewhere. If using several processors, each with one sensor, then you have a synchronization problem.

Why the cross post?

3

You can use multiple ESP32s. Each will handle 4–8 sensors. Use a Wi-Fi mesh or MQTT to send data to the central processing unit.

4

One way to obtain measurements from multiple sensors might be to use a single microcontroller for each sensor, collecting data at different times. However, this is not what I want. I want to obtain measurements from all sensors simultaneously.

5

Can I connect multiple sensors to a single ESP32 microcontroller? Does it have multiple connectors, or is there another way to do it?

7

Of course. Read them out one at a time. Are you reading the replies to your posts?

9

Yes, I’m reading the replies. Thanks! “One at a time” means it reads measurements from all attached sensors sequentially over time, right?

10

A good place to get started is to buy one sensor and follow the directions on the product page, the “Hookup guide” link.

11

In theory you can connect many sensors to a single ESP32. The problem is that many sensors have either a fixed i2c address, or a limited number of configurable addresses. For example, the TMF8821 has a fixed address at 0x41. You could connect two sensors to a single ESP32 using both i2c busses, but to get to more than 2 you either need multiple ESP32s (in which case you have a timing/synchronization problem to solve if you really need simultaneous readings), or you can use something like a mux (TCA9548A) to get up to 8 sensors with the same address on a single bus. The mux can be cascaded, but you can only read from one sensor at a time. So it really depends on your definition of “simultaneous”. If you mean reading several sensors within e.g. around 1 millisecond, you can cascade all you want. At 1 microsecond the problem gets a lot harder to solve.
The same question applies to orientation. If you need orientation accuracy with 5-10 degrees, then a BNO086 will probably be the simplest solution (and it’s used in robot vacuums already), but if you need more precision, then there’s a lot more work involved. There are some excellent libraries for IMUs, but they require careful calibration to work well, and their fusion algorithms can consume significant MCU CPU cycles.