Skip to content
MACT.ai
All work

Smart Telescope

Automated alignment, tracking and stacking in a consumer astronomy instrument.

2025Hardware · Embedded · Mobile · AI
Smart Telescope

Project Overview

An instrument that removes the hardest part of astrophotography. It solves its own position from the sky, drives two axes to the target, and stacks frames live while the user watches the image resolve on their phone.

Challenge

Plate solving, closed-loop motor control and live image stacking had to coexist on a battery-powered embedded platform, outdoors, in the cold, with only a phone as the interface.

Solution

We split the system across a real-time motion core and an application processor. Deterministic axis control stayed on an MCU while solving and stacking ran on Linux, connected by a framed serial protocol with defined failure behaviour.

Architecture

01

Hardware

A dual-processor architecture: an MCU for closed-loop motion, an application processor for imaging, with a custom mount driver board.

  • Two-axis closed-loop control with encoder feedback
  • Custom driver board with current sensing
  • Battery system with runtime telemetry
  • Sealed enclosure rated for outdoor overnight use
02

AI

Astrometric plate solving with a compact star-field matcher, plus a learned model for frame quality scoring before stacking.

  • Plate solve from a cold start in under 10 s
  • Learned frame-quality gate rejects trailed frames
  • Live stacking with rolling alignment
  • Automatic exposure and gain scheduling
03

Software

A mobile app over BLE and Wi-Fi Direct for setup, target selection and live view, with session capture stored on device.

  • iOS and Android from one codebase
  • BLE provisioning, Wi-Fi Direct for streaming
  • Live view with progressive stack preview
  • Session library with export

Result

  • Alignment to first tracked target in under 60 s
  • Sub-arcminute tracking error over long sessions
  • Full night of operation on one battery
  • Firmware and app shipped as a coordinated release train

Technology Stack

Hardware

MCU + MPUMotor DriversEncoders

AI

Plate SolvingFrame ScoringLive Stacking

Embedded

RTOSLinuxBLEWi-Fi Direct

Apps

iOSAndroidLive View