kepSmart is positioned in the energy management and sustainability space. Its unique and patented system makes it the only solution to offer savings into the 40% range consistently. This can only be achieved by giving the building owner/management the granular ability to control each room as its own zone. Not only measuring the temperature but actively being able to “control’ how much energy is used in that room to either heat or cool. 

kepSmart utilizes long range industrial wireless technology and intelligent battery management systems to offer non powered wireless sensors and controls to any building lasting a decade. Once the controls are in place, kepSmart‘s AI (artificial intelligence) engine takes over to make the building more efficient. 

Since we are the only solution to provide full control at each endpoint, our AI system can utilize a reinforcement learning (RL) model to the HVAC system of a multi-zone building to optimize boiler or chiller operation and maintain thermal comfort efficiently. Each zone is essentially a room, each assigned specific temperature weights based on usage or occupancy priorities. 

The RL agent controls the boiler or chiller and zone-level heating or cooling distribution, learning to balance energy consumption with comfort goals. It receives state information including past, present, and predicted temperature data, as well as external factors like wind direction and wind speed. Actions taken by the agent involve adjusting heating or cooling output and distribution strategies across zones. The reward function is designed to penalize energy overuse and comfort violations, encouraging the model to find an optimal policy over time. 

Training is often performed in simulation using historical building data and weather forecasts, allowing safe policy development before deployment. Once active, the RL system adapts continuously to dynamic environmental and occupancy changes, outperforming traditional control strategies.