A sensor-fused BIM-based intelligent control system for energy-efficient indoor environmental regulation - deep actor-critic reinforcement learning (DACRL)

Sep 30, 2026

Upload by Terk

A sensor-fused indoor environmental regulation system is built using deep fusion actor-critic reinforcement learning (DACRL) combined with multi-sensor fusion and Building Information Modeling (BIM). The four-layer cloud-edge-end architecture covers low-power perception via STM32L0 and LoRa sensors, spatio-temporal fusion with sliding-window weighting and EKF plus anomaly detection, and BIM-topology-based deep reinforcement learning decisions using Lieb–Thirring inequality as a spectral reward constraint. Real-time closed-loop control targets HVAC, lighting, and shading and yields up to 63.2% annual energy reduction versus rule-based control, while improving CO2 and thermal comfort metrics.

LanguageEnglish

CategoryTechnology

FormatPDF

Pages31

Reading Time1 h 18 min

AccessFree to download

Introduction

Background and motivation

Limitations of rule-based and MPC approaches

Bottlenecks for direct DRL application

Proposed system architecture

Four-layer cloud-edge-end design

Multi-sensor fusion and perception

Spatio-temporal synchronization and filtering

Anomaly detection pipeline

DACRL decision model

BIM semantic mapping and Lieb–Thirring spectral constraint

Execution and experiments

Real-time closed-loop control scenarios

Energy and comfort results

[www. nature.com/scientificreports](www. nature.com/scientificreports) OPEN A sensor-fused BIM-based ıntelligent control system for energy-efficient ındoor environmental regulation using deep actor-critic reinforcement learning (DACRL) Libin Tong The rising energy demand of the building sector, coupled with the need for high indoor environmental quality (IEQ), necessitates intelligent control strategies. However, traditional methods like rule-based control (RBC) lack adaptability, while model predictive control (MPC) suffers from model dependency. Hence, this study designed and implemented an intelligent indoor environmental control system based on multi-sensor fusion and Building Information Modeling (BIM). At the algorithmic level, this study employed a deep fusion actor-critic reinforcement learning (DACRL) algorithm to achieve energy-efficient dynamic environmental control. The system employs a four-layer architecture with cloud-edge-end collaboration. Layer 1: The perception layer utilizes a multi-modal sensor network built on an STM32L0 low-power micro-controller and an SX1276 LoRa communication module. This architecture enables distributed collection and pre-processing of parameters such as temperature, humidity, CO2, light, and occupancy. Layer 2: The fusion layer proposes a spatio-temporal synchronization mechanism for multi-source asynchronous data based on sliding window dynamic weighting and an extended Kalman filter (EKF). As an optimization, an isolation forest and one-class support vector machine (SVM) are combined to construct a multi-layer anomaly detection process. Layer 3: The decision layer establishes a deep reinforcement learning control model based on the topology of BIM semantic mapping. The key novelty of this model lies in its innovative embedding of the Lieb–Thirring (L–T) inequality, which describes the energy concentration characteristics of spatial fields, as a spectral constraint into both the reward function and the online action projection process of the Actor-Critic algorithm. This model innovatively embeds the Lieb–Thirring (L–T) inequality, which describes the energy concentration characteristics of spatial fields, as a spectral constraint into the reward function and online action projection process of the Actor-Critic algorithm. Layer 4: The execution layer implements real-time closed-loop control of terminal equipment such as HVAC, lighting, and shading through the standard BMS protocol. This study designed large-scale simulation experiments covering six core scenarios: offices, classrooms, residences, hospitals, shopping malls, and data centers. The experimental results demonstrate that the system’s annual energy consumption per unit area decreases by up to 63.2% compared to traditional rule-based control (RBC) strategies and by 23.5% compared to model predictive control (MPC) strategies. Atthe same time, several indoor comfort metrics improved significantly, with CO2 concentrations falling below 800 ppm 91.5% of the time and the thermal comfort (PMV) index remaining within ± 0.3 94.2% of the time. The proposed DACRL algorithm demonstrates excellent convergence and generalization capabilities. It achieves stable convergence numerically with an average of only 15,000 training steps. Performance loss in cross-building type transfer testing is only 4.9% . [www. nature.com/scientificreports/](www. nature.com/scientificreports/) Keywords Building ınformation modeling (BIM), Sensor fusion, Indoor environmental regulation, Energy efficiency, Deep actor-critic reinforcement learning (DACRL) College of Culture and Tourism, Fuzhou Polytechnic, Fuzhou 350108, Fujian, [China. email:](China. email:) tonglibin@sobeysca. com; [tlb@fvti.edu.cn](tlb@fvti.edu.cn) Energy consumption in the building sector is on the rise due to the acceleration of global urbanization and the growing demand for indoor environmental quality (IEQ) in contemporary buildings. This is now one of the main causes

FAQs

What problem does the proposed system target?

It targets intelligent control for energy-efficient indoor environmental regulation while maintaining high indoor environmental quality (IEQ) for building occupants.

How does the system handle multi-sensor data?

It uses a spatio-temporal synchronization mechanism for asynchronous data, combining sliding-window dynamic weighting with an extended Kalman filter, and adds multi-layer anomaly detection via isolation forest and one-class SVM.

What is the key novelty in the DACRL decision model?

The model embeds the Lieb–Thirring (L–T) inequality as a spectral constraint into both the reward function and the actor-critic online action projection process.

Sep 30, 2026

Upload by Terk