13 WSN Energy Management

13.1 Start With the Field Story

Spend Energy on the Decision, Not on Hidden Waiting

Picture a hill sensor that reports fire risk and relays messages for two nearby units. Its own report is cheap. Long listen periods, retries, and relay work may empty it first.

List sleep, wake, sense, compute, send, listen, relay, retry, and recover. Firmware is the built-in software that controls these states. A gateway is the unit that links the local sensor group to another network or service.

Measure current and time for a normal day and a blocked-link day. Turn off one route, add traffic, and replace a neighbour. Keep urgent safe action local and make low-energy doubt visible.

An average day cannot prove field life. Practitioner builds the energy record and controlled policy. Under the Hood explains relay burden, harvesting limits, and failure cases.

Start with the battery as a schedule, not a number on a datasheet. Every sample, wake period, radio listen window, relay burden, retransmission, and maintenance visit spends part of that schedule, so energy management is really a record of what the node is allowed to do and how often.

Phoebe’s Field Notes: Why The Relay’s Extra Current Costs More Than It Looks Like

The mathematical gist. A 1600 mAh pack at the chapter’s 50 µA leaf reference gives 32,000 h by linear division. A relay carrying six shares of traffic draws 300 µA, whose linear life is 5,333 h. Peukert’s correction, $t=t_{ref}(I_{ref}/I)^k$, reduces that to 4,876 h at $k=1.05$ or 3,116 h at $k=1.30$: chemistry changes how much extra life the higher current costs.

Math Bridge · guided foundationsWhy does relay current cost more than linear division predicts?Let Packet Pete turn Peukert's exponent into a bounded runtime penalty. Learn the maths

13.2 In 60 Seconds

WSN energy management is the evidence record that explains how a field network can keep sensing, forwarding, recovering, and being serviced for the promised monitoring period. It is not a single battery-life estimate. It is a claim about node states, traffic load, gateway paths, maintenance access, and what happens when conditions change.

The beginner mistake is to approve energy from a quiet-lab current number. A field node may spend energy on sensor warm-up, radio retries, receive windows, storage, clock repair, route changes, firmware updates, outage buffering, or recovery from weak links. The review has to preserve those states before it claims lifetime.

13.3 Energy Management Review Route

Pause the energy management review route discussion at the diagram in Figure 13.1. Comparing Monitoring claim against field exposes the evidence the next step would otherwise hide.

Notice how Figure 13.1 distinguishes Monitoring claim from field: the first defines what the design promises, but the second adds a distinct review condition. Add State budget, which adds a distinct review condition, to recover the whole system view. The figure’s conclusion is that wireless sensor network energy-management review route from monitoring claim through state budget, dominant drain, load path, energy policy, service margin, decision, known limit, and retest trigger. This conclusion belongs in energy management review route.

13.4 Beginner Examples

A soil-monitoring node that sleeps well can still fail if a distant gateway causes repeated retries. A bridge node with a good battery can still become the limiting node if it relays traffic for a cluster. A solar-assisted node still needs evidence for shaded periods, storage recovery, cold starts, and maintenance owner response. A longer sleep interval may save energy while making a control, alarm, or missing-data claim too weak.

13.5 Adapt Payload Detail as a Controlled Policy

Payload size trades information value against radio work. A very small report can omit trend or diagnostic evidence, while a large report occupies the channel longer, may need fragmentation, and costs more when interference forces a retry. The best choice changes with patient or process state, link quality, remaining energy, queue pressure, and the age of the evidence, so a fixed “smallest packet” rule is rarely an energy policy by itself.

One adaptive approach separates interpretation from sequential control. A fuzzy-inference layer turns imprecise inputs such as low/medium/high urgency, weak/acceptable link, or depleted/healthy energy into a bounded importance or risk score. A Markov decision process then represents the current state, permitted payload actions, transition uncertainty, and a reward that balances information delivered, energy used, delay, loss, and safety penalties. Its policy selects an action such as compact summary, normal report, or richer diagnostic payload, observes the result, and updates the next state.

For a wearable node, stable measurements and a weak battery may justify a compact periodic summary. A changing physiological signal can raise the importance score and select a richer payload even though it costs more energy; a severe alert should bypass optimization through a minimum safety payload and retry rule. The mechanism is valuable because it makes adaptation explicit, but it adds model, tuning, and verification cost. Fuzzy membership boundaries can hide subjective choices, an MDP learned in one environment can behave poorly after conditions drift, and optimizing average energy can starve rare events. Keep hard safety constraints outside the learned reward, log state/action/reward decisions, compare against a fixed-policy baseline, and reopen validation when sensing, radio, population, or operating conditions change.

13.6 Quick Check: Relay Load

13.7 Evidence Record

A practical energy record starts with the monitoring objective and then works backward through node states, traffic paths, service rules, and changes that reopen review. It should make the limiting condition visible instead of hiding it inside an average lifetime estimate.

13.8 Worked Reviews

Agricultural soil network: A soil network reports slow trends from battery nodes. Longer sleep intervals and local aggregation may be acceptable if missing readings stay visible and the dashboard decision tolerates delay. The record should include measured sensor warm-up, packet delivery, retry behavior, gateway outage buffering, service access after weather, and the lowest-margin node role.

Bridge monitoring: A bridge network may have a few relay nodes that carry traffic from hard-to-reach sensors. The practitioner record should review those relay roles separately from leaf nodes. If the relay fails first, the relevant lifetime is the relay service margin, not the average lifetime across all nodes.

13.9 Evidence Record Template

Evidence for Evidence Record Template starts in Figure 13.2. Look at WSN energy-management evidence record beside Claim scope before accepting WSN energy-management evidence record.

At Figure 13.2, WSN energy-management evidence record retains verification evidence; moving to Claim scope shows how it names a responsibility. The recovery that must hold label changes operating response. The pair WSN energy-management evidence record and recovery that must hold provides the review route for WSN energy-management evidence record. That makes Claim scope a checkable part of Evidence Record Template.

13.10 Knowledge Check: Evidence Record

LEACH makes relay burden visible in Figure 13.3 because the ordinary nodes and cluster heads follow different wake and transmit schedules.

Read Figure 13.3 at SETUP PHASE and STEADY STATE. Members sleep until slot reduces idle listening, but aggregate locally → one long uplink concentrates receive, compute, and radio cost at CH-A and CH-B; an energy record that counts only member transmissions will therefore approve the wrong lifetime model.

13.11 Where Energy Evidence Fails

Under the hood, energy failures often appear when a rare or hidden state becomes common. A weak link can turn transmit into repeated receive and retry work. A gateway outage can turn normal reporting into buffering and replay. A brownout loop can spend more energy recovering than sensing. A firmware update can change every state in the budget.

Harvesting needs the same discipline. Solar, vibration, thermal, or flow sources can extend service life when the measured source, storage element, load profile, quiet period, cold-start behavior, and owner response match the claim. They should not be used as a blanket “maintenance-free” label.

Average-current arithmetic is useful only after the state sequence is honest. The review should multiply each measured state by how long and how often it occurs, then add transition costs such as sensor warm-up, radio association, clock repair, storage erase, retry backoff, and acknowledgement waiting. A tiny active interval can still dominate if it repeats after every failed packet or every wake cycle.

Recovery states are easy to miss because they happen after something has already gone wrong. Brownout restart, queue rebuild, route rediscovery, buffered upload, and drained-storage recovery can turn a good sleep budget into a bad field budget. A WSN energy claim should therefore say what the node does after low voltage, weak link, gateway outage, or missed acknowledgement, not only what it does during normal reporting.

Role boundaries matter as much as state boundaries. A leaf node, relay, cluster head, and gateway-adjacent node may run the same firmware but spend energy in different proportions. The under-the-hood evidence should keep those roles separate so the first failing relay is not hidden behind a comfortable fleet average.

The under-the-hood rule is to keep the limiting evidence visible. If an optimization saves energy by adding latency, aggregation, local thresholds, extra gateways, or service work, the review should say which claim remains valid and which claim is no longer approved.

13.12 Knowledge Check: Harvesting Boundary

13.13 Summary

WSN energy management should protect the monitoring claim, not just quote a battery-life estimate. A reviewable record includes measured state budget, dominant drain, routing load, harvesting fit, service margin, owner, known limit, and retest trigger. Radio retries, relay roles, gateway behavior, event bursts, downlink commands, recovery loops, and firmware updates can dominate field energy. Average node lifetime can hide the first relay, cluster head, gateway-adjacent node, or event-heavy node that breaks the claim. Harvesting, duty cycling, local processing, topology control, and maintenance plans are useful only when their tradeoffs are explicit.

13.14 Key Takeaway

Approve WSN energy management only when the measured node states, traffic load, routing roles, service margin, harvesting assumptions, owner, and retest triggers support the same bounded monitoring claim.

13.15 See Also