What does this equation mean?

The third reason is that measurement itself has a cost and a failure mode. A current-sense shunt clamped around the supply rail is the standard way to verify what a design is actually drawing, as opposed to what its datasheet implies, and the discipline of doing this at multiple points in the duty cycle — sleep, wake-trigger evaluation, full inference, radio transmission — is what turns a power budget from arithmetic into an engineering artifact someone can defend. A budget built only from datasheet typical-current figures, without a bench measurement at each state, is analysis built on unverified vendor numbers stacked several layers deep; a single shunt measurement campaign, even a short one, converts most of that stack into fact. Here $f_i$ is the fraction of time the device spends in power state i (sleep, always-on monitoring, escalated inference, radio) and $P_i$ is the measured — not datasheet-typical — power draw in that state. This formulation is worth writing out because it exposes the one lever that dominates every other design decision: since $f_{\text{sleep}}$ is normally close to one and $P_{\text{sleep}}$ is normally the smallest term, the average power is overwhelmingly sensitive to the escalation rate — how often the cheap monitoring stage decides something interesting is happening — far more than to how efficient the expensive inference stage is once escalated. A system that escalates twice as often as necessary can double its average power even with a perfectly efficient inference core.