LAD-BNet: Lag-Aware Dual-Branch Networks for Real-Time Energy Forecasting on Edge Devices

TL;DR

This work tackles the challenge of real-time energy forecasting on edge devices by introducing LAD-BNet, a lag-aware dual-branch network that fuses a lag-centric dense path with a temporal convolutional network operating over the full sequence. The model is optimized for Google Coral Edge TPU, achieving ~18 ms inference and a 180 MB footprint, while delivering 14.49% MAPE at 1-hour horizon and robust multi-horizon performance up to 12 hours. Compared with LSTM and pure TCN baselines, LAD-BNet shows a 2.39–3.04 point improvement in MAPE and maintains strong R^2 across horizons, with a production-ready deployment pipeline including data processing, monitoring, and dashboard visualization. The work demonstrates substantial potential for industrial energy optimization, microgrid management, and edge AI, illustrating a practical and scalable path toward decentralized, low-latency energy forecasting.

Abstract

Real-time energy forecasting on edge devices represents a major challenge for smart grid optimization and intelligent buildings. We present LAD-BNet (Lag-Aware Dual-Branch Network), an innovative neural architecture optimized for edge inference with Google Coral TPU. Our hybrid approach combines a branch dedicated to explicit exploitation of temporal lags with a Temporal Convolutional Network (TCN) featuring dilated convolutions, enabling simultaneous capture of short and long-term dependencies. Tested on real energy consumption data with 10-minute temporal resolution, LAD-BNet achieves 14.49% MAPE at 1-hour horizon with only 18ms inference time on Edge TPU, representing an 8-12 x acceleration compared to CPU. The multi-scale architecture enables predictions up to 12 hours with controlled performance degradation. Our model demonstrates a 2.39% improvement over LSTM baselines and 3.04% over pure TCN architectures, while maintaining a 180MB memory footprint suitable for embedded device constraints. These results pave the way for industrial applications in real-time energy optimization, demand management, and operational planning.

Figures (8)

• Figure 1: LAD-BNet dual-branch architecture. The Lag Branch (left, purple) processes the last 24 timesteps through dense layers to capture short-term dependencies. The TCN Branch (right, blue) processes the full sequence via dilated convolutions to capture long-term patterns. The Fusion Module (center, green) combines both representations through dense layers for multi-horizon predictions (1--12 hours). The architecture comprises $\sim$245,000 parameters with a 271.6 KB quantized model size (int8), enabling efficient edge deployment.
• Figure 2: Feature engineering pipeline transforming raw sensor data into the 27-dimensional feature vector. Six feature categories are derived: (1) Temporal cyclical features; (2) Meteorological features; (3) Contextual binary features; (4) Temporal lag features; (5) Rolling statistics; (6) Interaction term. All features are normalized to [0,1] via MinMaxScaler before model input.
• Figure 3: Multi-horizon forecasting performance comparison. MAPE evolution across prediction horizons (1h, 2h, 4h, 8h, 12h) for LAD-BNet versus baseline methods. The horizontal dashed line at 15% indicates the target objective threshold. LAD-BNet maintains MAPE below 15% up to 4-hour horizon.
• Figure 4: Baseline comparison at 1-hour prediction horizon. Horizontal bar chart showing MAPE (%) for each model. LAD-BNet is the only model meeting the 15% target threshold.
• Figure 5: One-week test period showing actual energy consumption versus LAD-BNet predictions. Blue solid line: Actual measured consumption (kW). Orange solid line: LAD-BNet 1-hour ahead predictions. Gray shaded region: 95% confidence interval. The model accurately captures daily consumption patterns.