ESP32-powered Electrical Impedance Tomography Toolkit by MIT

Reporting from Shanghai, China
Nov 22, 2021

A team of engineers from the world-renowned Massachusetts Institute of Technology published a paper about an ESP32-based EIT-kit for health and motion sensing.

A team of engineers from the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology (MIT) presented a paper on Human-Computer Interaction (HCI) at last month’s ACM Symposium on User Interface Software and Technology (UIST). In their paper, Junyi Zhu, Jackson C. Snowden, Joshua Verdejo, Emily Chen, Paul Zhang, Hamid Ghaednia, Joseph H. Schwab, and Stefanie Mueller presented the EIT-kit, an ESP32-based electrical impedance tomography toolkit for designing and constructing health- and motion-sensing devices.

Electrical Impedance Tomography (EIT) is an imaging technique that measures the conductivity, permittivity, and impedance of a subject. In the past, EIT-sensing required large, expensive hardware setups, as well as complicated image reconstruction algorithms. However, with the recent development of low-cost electronics and the availability of open-source EIT image reconstruction libraries, such as EIDORS, EIT-sensing has become accessible to Human-Computer-Interaction (HCI) researchers who have used it for touch sensing, tactile sensing, and hand-gesture recognition. Such advances have made EIT-sensing more portable, thus demonstrating the great potential of low-cost EIT technology and its benefits for the health-sensing domain, especially in sports medicine and home care.

However, the expertise required for designing custom EIT devices is still high. To create an EIT device, users have to design the form factor of the device first, in order to ensure continuous contact between the electrodes and the subject, which can be different each time, depending on the measurement location and electrode distribution.

In the above-mentioned MIT paper, the authors present the EIT-kit, an electrical impedance tomography toolkit that supports users across the different stages of EIT device development. The EIT-kit consists of:

  • a 3D editor for customizing the form factor of the measurement setup and the electrode distribution;
  • a customized, ESP32-based, EIT-sensing motherboard that supports different measurement setups, i.e., two- or four-terminal setups, up to 64 electrodes, and single- to quadruple-electrode arrays simultaneously. The EIT-sensing motherboard also provides adjustable AC injecting current to improve the quality of the signals;
  • a microcontroller library that automates the calibration of the signals, and facilitates data collection;
  • an image reconstruction API for mobile devices that can be used for interpolating and then visualizing the data.

The authors explain the development of their EIT-kit with a formative user study. Then, they demonstrate the capability of the EIT-kit to support various interactive devices that focus on health and motion sensing (i.e., a muscle monitor for physical rehabilitation, a wearable hand-gesture recognizer, armbands for non-intrusive driving detection). Finally, they conduct a technical evaluation of the data fidelity of their EIT measurements.

The EIT-kit

The EIT-kit supports users across the different stages of EIT device development, by offering a 3D editor for creating custom measurement setups for different measuring locations (e.g, wrist, thigh) and sensing resolutions (number of electrodes, electrode distribution). In the EIT data measuring stage, the EIT-kit includes an ESP32-based EIT-sensing motherboard, and a sensing library (Arduino-based) for acquiring data from the board. In the final stage, where users have to interpolate the data and visualize them, the EIT-kit assists users with an image reconstruction API for mobile devices (i.e. iOS devices) that is capable of 2D and 3D visualizations, both on screen and in augmented reality (AR). By providing these building blocks that are essential for EIT device development, the EIT-kit facilitates the creation of custom EIT-sensing applications.

EIT-Sensing Motherboard

EIT-kit’s sensing motherboard is based on ESP32, and is capable of automating the EIT signal calibration and measurement for different electrode configurations.

EIT-sensing motherboard: (a) top view, (b) bottom view, (c) with two stacked up mux boards. Photo source:

More specifically, the motherboard consists of the main sensing board, which is responsible for injecting the AC signal and measuring the resulting voltage output, and up to two multiplexer boards stacked on top of each other, directing the signal to individual electrodes. The main sensing board has three components:

  • a current drive circuit for injecting the AC signal;
  • a voltage measurement circuit for measuring the voltage output from the current drive, and
  • a control circuit with an ESP32 microcontroller.

(a) Top and (b) bottom view of the EIT-sensing board highlighting the parts that make up the current drive, voltage measurements, and control circuit. Photo source:

Both the current drive- and voltage-measurement circuits are controlled by the ESP32 microcontroller via the SPI channels and the GPIO pins. In order to have more direct control at faster frequencies, the MIT engineers implemented the control circuit via two separate SPI buses (HSPI & VSPI). The first SPI bus is used to control the signal generator and digital rheostat. The second SPI bus is used for the IO expander (MCP23S17) that drives the chip-select pins of the multiplexers, as well as various other digital inputs. The ADC converter output is routed directly to the ESP32 GPIO pins, and sampled at 20MHz.

To get all the details about the ESP32-powered EIT-kit, please check out the MIT paper here. Currently, EIT-kit is expected to facilitate the work of interactive-device designers who focus on health and motion sensing. Since interactive-device designers often use 3D modeling and 3D printing, in order to build new devices and write code to program them, EIT-kit’s 3D editor plugin as well as the microcontroller and visualization APIs match the needs, as well as the scientific background of the above-mentioned target group. To extend EIT-kit’s audience to users who are less experienced in programming, the MIT engineers will extend EIT-kit with pre-built applications as part of their future work.

EIT-kit supports users in creating a variety of EIT-sensing devices (a) and visualizing the resulting data (b). EIT-kit provides a 3D editor plugin (c), a sensing motherboard based on ESP32 (d), and a microcontroller library, as well as an image reconstruction API. Photo source:

Share this article
  • LinkedIn
  • 微信


Reuse this content