CSA: Directed Retry Can Alter Mobile Phone's Location

CSA: Directed Retry Can Alter Mobile Phone's Location

Class│Value│
0 0 0│1 1 0 1│ Directed Retry

The accuracy or inaccuracy of cell site analysis testing measurements largely depends upon what has been considered and there are indeed many points to consider. One handover (HO) procedure, if it is included within a mobile network's radio-availability and traffic-flow arsenal, is called Directed Retry (DR). The GSM and 3GPP standards refers to this procedure.



What is Directed Retry (DR)?
Directed retry has adjustable parameters in order to define thresholds that once passed can trigger DR. When DR is set as Not Use it is inactive. Once set to Use the default value is set until the parameter is adjusted. That is to say a 'value' that is set as default can be modified in response to condtions eg quality of service (QoS) or traffic observations. A manufacturer of the say the BSS may provide recommended values, but it might be the OMC-R or BSS engineering team may require to make their own determination about values for internal or external handover procedures.

Use DR enables for example the BSS to move a mobile phone's communications to another cell (Mast or sector of a Mast) prior to call set up. That can be for an outgoing or incoming communication.

DR may be triggered by, for instance, due to 'congestion' and therefore may require internal or external handover procedures to combat that traffic condition. An outcome is that a mobile phone that receives service from the current serving cell (maybe the Mast is seen as closer to the mobile phone as well) is handed over to a cell that originates from a Mast that could be eg:

- some distance from the mobile phone's actual location
- coverage from a adjacent Mast in an area
- etc

This is one of many radio cases that when conducting radio test measurments a 2G/3G passive radio detection device and its readings may not record the appropriate network messages and thus mis-inform their users attempts in assessing a mobile phone's general location when conducting cell site analysis, as the device's readings may be incomplete. The Cell ID obtained from a call detail record (CDR) can only reflect the antenna identities on a fixed-positioned Mast and that a mobile phone has had its communications routed to and from the network using a particular Mast (so to speak). It doesn't automatically follow that the Cell ID confirms the general local area in which the mobile phone was actually located without certain radio data and other necessary checks being made.

Cell Site Analysis - Criminal Case

Cell Site Analysis - Criminal Case

IN HER MAJESTY'S COURT OF APPEAL IN NORTHERN IRELAND

Neutral Citation no. [2008] NICA 5

THE QUEEN v RICHARD DAVID McCARTAN and BARRY DAVID SKINNER.

Appeal Court: Kerr LCJ, Campbell LJ and Morgan J

http://www.courtsni.gov.uk/NR/rdonlyres/4F39A7FC-5350-4E4B-BB76-FA8DC4BA08D6/0/j_j_KER7057Final.htm


It is not proven with cell site analysis that evidence obtained relating to mobile call records, the Masts and their details associated with handling the calls and conducting radio test surveys combined together form a precise indication of the geographical location of a mobile telephone at the material time. Therefore, it is unlikely an alleged opinion that it is consistent to place a mobile telephone at a particular location holds weight. That cell site analysis might generally indicate the possible movement of a mobile telephone in relation to Masts that it used and no more.

Whilst the Appeal Case for the appellants (McCartan and Skinner) did not succeed it was not because cell site analysis was able to pinpoint the location of the appellants' mobile telephones, which it is unable to do, but the appellants' failure to identity what was happening with their mobile telephones at the material times. The following commentary from this Appeal Case is worth reading.

'The cell site analysis

[37] An expert in this field, David Sanderson, gave evidence, submitted a report on the subject and made a Power Point presentation to the court. Mr Fee suggested that Mr Sanderson’s evidence established that (a) call mapping was carried out in respect of telephone 008 (attributed to Mr McCartan) and telephone 301 (attributed to Mr Skinner); (b) call mapping provides an approximate indication of where a mobile phone was when a call was made; (c) call mapping is used primarily to show the likely movement of a cellular phone with a general indication of location; (d) the cellular phone may be anywhere within the best serving coverage of the sector; (e) there is boundary overlap between sectors served by different masts; (f) the extent of the best serving coverage sector is unknown; (g) the indication of movement of the person using the telephone is no more than consistent with a suggested pattern of progress – it cannot establish positively that the user of the telephone in fact moved in the mooted direction; (h) calls 15, 16 and 17 from 008 at 20.53: 20.55: 20.55 were made when the telephone was not in the vicinity of Euston Street.

[38] On the basis of this evidence Mr Fee argued that the cell site analysis did no more than provide a very imprecise indication of general movements. In the absence of a capacity to identify a precise location of the telephone within the area of best coverage it was not possible to provide any precise indication of travel direction. We are disposed to accept that Mr Fee has correctly identified the limitations that apply to the cell site analysis. As he has acknowledged, however, it does provide a general indication of the movements of the users of the telephones. If the appellants followed a different pattern of movements on the day in question, it was open to them to give evidence about those movements and to explain where they had been at the critical time in relation to the shooting. Their failure to do so, combined with the general indication that the cell site analysis provided, constituted strong evidence against them that their movements were as indicated by that analysis. In Mr McCartan’s case, his inquiry of the interviewing police officers as to how precisely he could be tied down to a particular area was also, in our judgement, extremely telling.'

The comments of the Appeal Court in para [38] suggests the appellants' own lack of self-preservation to state what their mobile telephones were doing and the reasons for going in a direction following the line of Masts used for handling calls, undermined the cell site analysis element of their appeal.

GSM Radio Test Measurements - Non-Dominance

GSM Radio Test Measurements - Non-Dominance

Screen Image 1

Readers and cell site analysis students will recall the thread GSM Radio Test Measurements, below. In that thread the discussion related to possible anomalies and interpretation regarding radio test measurements. In this thread I want to highlight another radio anomaly termed non-dominance that may occur.

Generally speaking, non-dominance occurs when radio coverage from two or more radio sources are equally aligned so that they all become dominant. The rare event in Screen Image 1 displays coverage quality and signal strengths of four radio sources are all equal to each other, in a confined area, where the receiver is in the idle mode and at ground level. Attempting to define that one particular Mast would be used to make or receive a mobile call in these circumstances might be difficult. The problems interpreting outcomes due to non-dominance could be numerous. For instance:

i) Who is to say that the receiver camped on a Mast's coverage shown by ACT would actually use that Mast? The coverage at NC2, NC3 and NC4 are equally as likely to be candidates to carry a mobile call.
ii) What happens in the case that the Mast's coverage upon which the receiver is camped at ACT the Mast is not actually best placed or line of sight with the receiver and in other circumstance would not be considered at first instance to serve the geographical area? The receiver under these conditions could be placed in an entirely different geographical area simply from summarising the Mast details shown in the mobile call records.
iii) When call setup (OACSU) takes place and a connection made it could be possible a mobile ‘phone starts a connection for less than a second on the Mast at ACT but forced into hard handover immediately thereafter thus shifts the call to another Mast (say, NC4)? It could be possible the Mast at ACT is shown as the Mast in the call records, which could be rather misleading when attempting to consider geographical location of the mobile telephone when compared to a particular mobile call.
iv) How can the radio examiner, from looking at historical records, correspond radio test measurements against the outward set of event elements recorded in call records? Non-dominance may mean a series of calls being made at one location handled by numerous Masts – thus may amount to a suggestion from analysis of the Masts usage from the call records that the mobile is on the move when in fact it is rather confined to an area. Non-dominance may result in a high number of short duration mobile calls shown in the call records. The short duration calls may not be due to the user terminating the calls, but rather the network dropping the calls.

I introduced, I believe, GSM non-dominance as an important radio anomaly to be considered regarding evidence of radio test measurements into the first criminal case in the UK back in 2003 at the Central Criminal Court (Old Bailey). GSM non-dominance was identified and shown in my evidence. Three other experts appearing for the prosecution or other defendants received my report. Acceptance of my results and findings were agreed and no challenges were made. However, I should point out that the discovery that non-dominance can occur for radio coverage was not my invention but arose from radio propagation studies by mobile network radio engineers endeavouring to agree anomalies for radio principles. The non-dominance principle had already been adopted for the Tetra standards and is accepted in GSM radio parlances as a noteworthy event. Therefore, I was fairly well armed with other independent findings.

The purpose of this discussion thread is to provide yet another example of radio conditions that can prevail when conducting GSM radio test measurements. That simply analysing call records may only present half a picture – a trompe l’oiel ( a lie to the eye) if you will.