Cleaning robots are appearing on the streets of Russian cities, and with them comes a new problem: nobody can check the quality of the work by eye any more. No engineer walks beside the machine, no dispatcher sees how much water went onto a particular stretch of pavement, and compliance with the schedule rests on trusting the automation. Avtonomika solved this together with Galileosky: telematics terminals collect location, water consumption and vehicle status data, while Easy Logic algorithms turn that stream into actionable events inside the robot itself.
Why manual cleaning no longer scales
Municipal services today operate under hard constraints: staff shortages, rising consumable costs, and almost no way to verify the quality of low-skilled manual labour. A single cleaning robot covers the output of up to five street sweepers, while its upkeep takes about half of a skilled position — an engineer, designer or programmer. That exchange changes the cost structure: instead of staff turnover you get a maintained asset with a predictable service life.
The second driver is environmental. Diesel-powered cleaning equipment produces emissions and a constant noise floor that residents complain about regularly. An electric robot cuts both: the project estimates up to 50 tonnes of CO₂ per year per unit and up to 40% less noise. The third motive is digitalisation of the utilities sector, which has long lagged behind on innovation. Autonomous machines pull the regulatory framework, charging infrastructure and new accounting requirements along with them.
What Pixel can do and why it needs a monitoring system
Pixel is an autonomous machine with interchangeable attachments for all-season cleaning and a computer vision system. In winter it handles snow and de-icing agents; in summer, dust, sand and pressure washing. The attachments change, and so does the set of parameters to watch: water consumption during washing, brush operation, running hours per module.
The task the customer set
Avtonomika needed a monitoring and dispatch system that gathers data on how the robots perform. The main risk was stated plainly: robot data must not be lost in an emergency, a technical failure or any other incident. If a robot stalls on its route, is involved in a collision or loses connectivity, the record of its work must still be in the operator's hands.
Why an ordinary tracker is not enough here
A robot is not a car with a driver who can explain what happened. There is no person to confirm that cleaning took place, and no paper waybill. Telematics therefore becomes the only witness: it records the track, the pace of work, water consumption and machine behaviour at every moment. An error in the data equals an error in accepting the job.
How the Galileosky-based solution works
The project uses Galileosky 10 and Galileosky 7x terminals. The first is the flagship model with an extended interface set — RS-485, RS-232, 1-Wire, up to ten inputs and four outputs for controlling actuators — plus CAN scanner support. The second is a versatile GPS/GLONASS terminal for flexible management of fleets and stationary assets. Both work with two CAN buses and support J1939, J1979, FMS, CANopen and NMEA-2000 protocols.
Precise positioning and telemetry collection
The terminal determines the robot's coordinates while collecting status data. The monitoring system shows where the machine is, how fast it moves, which mechanisms are engaged and how much resource is being consumed. For cleaning this matters: the track shows the route actually covered rather than the one in the plan, and it makes it easy to match the treated area against water usage.
Data duplication as a black box
Reserve storage is the key function in this project. The terminal writes data to its own memory and uploads it to the server once connectivity returns. Even if the robot is involved in an incident and loses power or network, telemetry from the last hours of operation remains available. This is exactly what closes the customer's requirement: Pixel data is not lost in emergencies.
Easy Logic algorithms inside the robot
Easy Logic is a Galileosky technology that lets you build algorithms from ready-made blocks without programming skills. In this project the scripts analyse location and acceleration data and pass the result onto the robot's CAN bus. The result is feedback in both directions: not only "we see what the machine does", but also "the machine knows what is happening to it". The advantage is that the logic can be adapted to a new cleaning scenario without rewriting firmware or waiting for a vendor update.
What changed after deployment
The project delivered measurable results. The need for low-skilled staff dropped: one robot replaces up to five workers. CO₂ emissions fell by 50 tonnes per year per unit and noise levels by 40%. Cleaning became more transparent thanks to automated compliance monitoring: the dispatcher sees that a route was actually covered and how much resource was used, rather than a report based on the operator's word. Predictive analytics built on telemetry data reduced unplanned repairs by 15–20%.
Scaling up and the next step
The solution is not a pilot on a single site: it is scaling in Moscow, St Petersburg and other regions. The logic is simple — the more robots in the fleet, the more valuable a single dispatch system that shows every unit and every cleaning job. Robotic cleaning is moving from experiment to routine municipal service, which means the accounting requirements will only grow.
Contact: sales@galileosky.com