What Is Campus Energy Saving?

Device-Level Energy Saving

Device-level energy saving refers to defining different power modes for each device depending on its load. For example, idle devices can enter sleep or standby mode, while lightly-loaded devices can switch to low-power mode.

In campus networks, APs are the most numerous devices and therefore the primary focus of energy-saving strategies. When idle, an AP enters sleep mode, with nonessential modules shut down and minimal power maintained only for critical components such as the CPU. This reduces energy consumption by 80–90% on a single AP and allows the AP to resume operation within one minute. In addition, since APs are typically powered by PoE switches, further savings can be achieved by disabling switch ports to cut off power supply completely. In this case, APs consume no energy until reactivated, at which point the corresponding switch ports are re-enabled.

Energy Consumption Visualization and AI-Driven Optimization Based on Tidal Traffic Patterns

Traffic on campus networks exhibits distinct tidal patterns. For example, as shown in the following figure, traffic in a teaching building rises sharply at around 7 a.m., peaks in the morning and afternoon, and drops to nearly zero after midnight. Energy-saving measures can be applied to APs during off-peak hours based on such tidal traffic patterns to maximize energy efficiency across the entire network.

Tidal traffic patterns of a university teaching building

Currently, the common energy-saving approach is to disable Wi-Fi services during predefined time periods, but this has two drawbacks:

• Uncertain time windows: It is difficult to determine whether users require access during the shutdown period, which may result in poor user experience.

• Uncertain AP selection: As more and more IoT devices connect through Wi-Fi or AP-integrated IoT base stations, directly disabling APs can interrupt IoT services.

To address these limitations, AI can be leveraged to analyze network data and intelligently recommend energy-saving strategies for APs. This guarantees user connectivity and IoT services while minimizing overall energy consumption, achieving both energy efficiency and optimal user experience.

CSI Sensing-based Energy Saving

Beyond network devices, energy saving for IoT devices is also critical. Significant energy savings can be achieved by adjusting devices in real time based on human movement, for example, automatically turning lights and smart screens on and off, and regulating air-conditioning temperature.

CSI sensing, an advanced WLAN technology, enables statistical analysis of human activity within physical spaces while ensuring information security. With widespread Wi-Fi coverage, CSI is an effective approach to intelligent campus energy saving.

Like network devices, electrical equipment also exhibits tidal patterns in traffic. By automatically adapting equipment behavior to occupancy, such as adjusting lighting brightness or air-conditioning temperature, the campus energy-saving solution can substantially reduce power consumption. However, unlike meeting rooms, indoor open areas lack clear boundaries, making CSI data from a single AP insufficient for direct device control. To address this, a third-party system is used to aggregate sensing results from multiple APs. Energy-saving measures are triggered only when no occupancy is detected for more than 10 minutes (15 minutes recommended). The following figure shows how a campus network leverages CSI sensing to save energy for buildings.

CSI sensing-based building energy saving

Energy-Saving Window Prediction

A proper network energy-saving window requires predictable tidal traffic patterns. Low-traffic periods must meet the following conditions:

Light-load period:

• The number of terminals is significantly lower than during peak hours.
• The number of terminals remains relatively stable during this period, though some devices (such as office laptops, stationary mobile terminals, and always-on IoT devices) may stay connected.

Load decline period:

• The number of terminals drops from peak levels toward light-load levels without a significant rebound.
• The duration should be at least two hours.

To identify proper time windows, NCE-CampusInsight analyzes historical terminal data reported by APs to generate time-based network tidal sequences. It verifies sequence stability to determine whether traffic exhibits regular patterns. If irregular, no energy-saving window is recommended. If regular, the anomaly detection algorithm is applied to remove sequence disturbances, and the time series prediction algorithm is applied to forecast future tidal sequences. Based on these results, NCE-CampusInsight recommends energy-saving windows that align with tidal troughs.

Tidal patterns of the user access count

Intelligent IoT AP Identification

IoT APs typically operate 24/7 to provide continuous Wi-Fi data transmission for IoT terminals, making them unsuitable to be powered off for energy saving. NCE-CampusInsight can intelligently identity IoT APs based on AP-reported data. Once an AP is recognized as IoT-specific, it will not be recommended for energy-saving measures.

• IoT APs often integrate radio capabilities such as RFID, BLE, Zigbee, or UWB to provide radio connectivity for IoT terminals.
• They can identify Wi-Fi-connected IoT terminals based on terminal fingerprints and behavior characteristics.

Intelligent IoT AP identification

CSI Sensing

CSI sensing involves algorithm-based detection and identification, performed independently by APs acting as both the transmitter and receiver, without coordination with neighboring APs. For details, see CSI Sensing.