A I – and СV – enabled autopilot for autonomous electric mobile platforms
One approach, many solutions
Using our own development, the Mars 500 robotic universal platform, as a testbed, the MarsMe team is developing a modular autopilot based on the use of AI algorithms, deep machine learning via multilayer neural networks.
We use a budget-friendly, proven approach to our solutions by using computer vision as the main tool for information capture and processing. This allows for automation with the help of inexpensive cameras and sensors.
The autopilot’s versatility allows it to be used to robotize almost any electric mobile platform that can use attachable or trailed equipment.
The autopilot consists of a base unit, specialized modules, and a cloud platform.
The basic autopilot unit allows you to implement the following functionality:
- Movement and navigation on a previously set route.
- Orientation and navigation between obstacles
- Territory recognition (highways, footpaths, sidewalks, industrial areas)
- People, animals, and objects detection.
- Built-in safety functions.
Autopilot specialized modules allows you to perform the following tasks:
- Road surface cleaning;
- Snow cleaning;
- Sand, granite, and salt sprinkler for de-icing in winter
- Solar panel cleaning;
- Road marking application;
- Treatment of land plots;
- Dispensing animal feed;
- Manure removal;
- Fodder mover;
- Digging trenches and drilling small wells;
- Cargo hauling;
- Firefighting and other rescue work;
- Leaves removal;
- Grass mowing and other work related to landscaping;
- Monitoring of the territory and facilities security;
- Data collection in agriculture and geological prospecting works.
The implementation of these autopilot modules in the production will go as demand exists.
The cloud-based autopilot’ control system platform allows:
- Carry out its learning remotely, to solve non-standard tasks, thereby extending the functionality of the equipment attached.
- Remotely update the autopilot software.
- Remotely control the operating modes and location of the machine.
- Use additional computing resources, if necessary.
With our API, which allows other developers to connect to our cloud platform, the autopilot can be integrated into smart cities and smart farms.
We are currently interested in finding an investor/partner to improve existing and further developments and their subsequent commercialization