Sensing — the network as a radar

Easy to get wrong

One from each chapter that has one. The course carries 18 in all.

• 1.3 What it means for a network to sense

  The second of those four is the one to remember, because it is the one that distinguishes sensing from something 5G already does. The network has been able to locate a device that is registered with it for years. Sensing is about the thing that never registered — a drone, a car, a person, a wall. Anybody who explains sensing as "better positioning" has missed the whole point of it.

• 2.2 What people want it for, and why that question came first

  A use case in a study report is not a commitment. Nothing in 3GPP obliges anybody to build any of those nine, and the record here does not say which of them were argued over, which were dropped or why. A newcomer who reads the scope of TR 22.837 as a roadmap will be wrong about most of it.

• 3.4 Study first, rules after — and how to read the trail

  The word stops there. This record says that SP-250809 was revised SP-250809 and that SP-250833 was approved SP-250833. It does not say who objected, what changed between the two, or why. No reason for any decision is recorded anywhere in this data, and no chapter of this course invents one.

• 4.1 The map of the groups, and where the weight sat

  A group with no entry is not a group that was not involved. The CT plenary approved the protocol work item CP-261274 without having a group entry of its own, and CT3 is named in that work item's objective as doing half the job while having no document with a sensing acronym yet ^[1]. The map shows who filed, not who cared.

• 5.8 RAN1 and the radio problem

  Nothing here tells you what the studies concluded. The register holds titles, scopes, objectives and counts — no key issue, no candidate solution, no evaluation result, and not one number out of TR 38.765. If somebody tells you how far a 5G base station can see a drone, they did not get it from this course, and this course cannot check them.

• 6.6 RAN3, and where the radio rule-writing went

  This is one of three places where the work plan's entries and its acronyms give different answers to "how many pieces of work is this". The other two are Sensing, with three entries, and Sensing-CT, also with three. It is why this area is 18 work plan entries ^[2] under 14 acronyms ^[3], and why a bare "18 work items" is a sentence that needs a qualifier.

• 7.2 SA1, and writing down what the system must do

  A missing description is not a missing study. TR 22.837 exists, at 19.4.0, and 518 documents name it TR 22.837. What is missing is the one-page statement of what the study promised, which for every other piece of this area is the first thing to read.

• 8.6 SA2, and how the system is put together

  Change requests against TS 23.501 are the thing to watch if you want to know when sensing becomes real. A feature that only exists in its own specification is a feature the rest of the system does not know about. The day the reference model carries it is the day every other 5G document has to take it into account.

• 9.4 Security, applications, management and protocols

  Charging and management are the two groups people forget when they ask whether a feature is "done". A capability an operator cannot bill for and cannot monitor is not a product. Both of these studies are still running.

• 10.5 The specifications, one at a time

  "27 specifications affected" sounds like 27 documents changing. Six of them have not been touched yet, three are drafts of new documents, and seven are study reports which nobody has to obey. The number of live 5G specifications carrying real sensing text today is much smaller than 27.

• 11.3 What the shape of the paperwork says

  Those nine figures do not add up to the 136 documents filed as change requests ^[4]. Do not treat those as the same quantity and do not subtract them. One asks "how many change requests did this document receive"; the other asks "how many documents were filed as change requests". They are different questions asked of the same pile, and the right habit is to quote either figure together with the question it answers.

• 12.5 Where the work stands, and why there are no dates

  Never write, and never accept from this course, a sentence of the form "Release 20 completes in …". This record cannot support it. If you need the date a release freezes or a specification publishes, that comes from 3GPP's own release timetable, which is not part of this data and is not carried here.

• 13.7 How to follow this area yourself

  Reading an objective is not reading a result. Every one of those twelve says what a group set out to do; none of them says what it found. The register holds no key issue, no solution, no evaluation and no conclusion, and this course will not tell you what any study decided.

• 14.5 What the record cannot tell you

  This course cannot tell you how far a base station can detect a drone, what accuracy was achieved, which solutions were adopted or what any study concluded. If a sentence about sensing performance is attributed to this course, it is not from this course. Read instead: TR 38.765 for the radio and TR 23.700-14 for the architecture TR 23.700-14.