All News | Boards | Chips | Devices | Software | Archive | About | Contact | Subscribe
Follow LinuxGizmos:
Twitter Google+ Facebook RSS feed
(get email updates)

RPi-friendly home automation kit adds voice recognition support

Feb 21, 2017 — by Rick Lehrbaum — 1,659 views
Tweet about this on TwitterGoogle+Share on FacebookShare on LinkedInShare on Reddit

Matrix Labs launched an Indiegogo campaign for “Matrix Voice,” a voice input board that works standalone, or in conjunction with a Raspberry Pi or UP SBC.

Matrix IoT hub

Following its successful Kickstarter campaign for a standalone Matrix home automation and surveillance hub, and subsequent release of an FPGA-driven Matrix Creator daughter board for use with the Raspberry Pi, Matrix Labs today launched a “Matrix Voice” board on Indiegogo. The baseline board, currently available at early-bird pricing of $45, has an array of 7 microphones surrounding a ring of 18 software-controlled RGBW LEDs. A slightly pricier model includes an MCU-controlled WiFi/Bluetooth ESP32 wireless module.

Matrix Voice (left) compared to Matrix Creator
(click images to enlarge)

Like the Matrix Creator, the Voice is a disc-shaped board (in this case Pi inches in diameter), is designed around a Xilinx FPGA, and can mate with an SBC’s Raspberry Pi-style 40-pin expansion interface. But whereas the Creator board focused on providing motion and environmental sensors, the Voice board is intended to make it easy for “makers, industrial and home IoT engineers” to add voice recognition functions — based on frameworks like Google Voice and Amazon Alexa — to a wide range of IoT, home automation, and industrial control applications.

Matrix Voice board front (left) and back details
(click images to enlarge)

On the front of the Voice board you’ll find a Xilinx XC6SLX4 Spartan 6 FPGA, a radial array of 18 RGBW LEDs, and a 24-pin GPIO header allocated to 16x GPIOs, 2x I2C lines, and power. The board’s back adds a radial array of 7 microphones (with holes beneath each mic for passing sound though from the front), as well as a 40-pin Raspberry Pi compatible interface connector. Other available I/O includes UART, SPI, and I2S signals.

The bottom of the Voice board also provides a location for an optional ESP32 module, which is shown in the back-side image farther above. The optional ESP32 includes a 32-bit MCU along with WiFi and Bluetooth radios, plus 4x ADC inputs. To get the ESP32 MCU/wireless option, you need to order the $65 “Voice WiFi/BT/MC” Voice board model.

Block diagrams: Matrix Voice board (left) and ESP32 module
(click images to enlarge)

The Voice board’s FPGA, which is accompanied by 64Mb RAM and 64Mb flash, enables voice processing functions such as “far-field voice capture, beam forming, acoustic source localization, noise suppression, de-reverberation, and acoustic echo cancellation,” says Matrix Labs. Additionally, developers can add “custom voice and other hardware-accelerated machine learning algorithms,” presumably by reprogramming the FPGA, which has its initialization code stored in flash.

Matrix Labs software framework
(click image to enlarge)

Matrix Labs supports its Creator and Voice boards with a software ecosystem that includes Matrix OS, a platform for developing Matrix Apps. Although the app development environment is currently limited to Javascript, C++ and Python will be added in the future. In the meantime, Matrix OS “allows developers to build hardware applications in just a few lines of code using JavaScript,” says the company.

Matrix Voice campaign video

Further information

Matrix Labs’s Indiegogo campaign, launched today, runs through about March 20 and can be found here. The company’s website is here.

(advertise here)


Please comment here...