ART releases the results of a second survey to assess coverage by mobile phone networks

Paris, 13 December 2001

France's Telecommunications Regulatory Authority (ART) recently conducted a second survey to assess coverage by mobile networks in sixty cantons. This follows on from an initial survey of forty cantons carried out last June. By the end of 2001 therefore, ART will have tested coverage in a total of one hundred cantons.

 

A key area of ART's activities, in accordance with one of the objectives assigned by the 1996 Telecoms Act, is to address issues of regional development. It is therefore natural that ART should contribute to the discussion currently in progress on nationwide mobile network coverage.

 

The "field" phase of the survey was conducted by Directique, a specialised consultancy, over a six week period beginning in October.

 

The method used for the second survey differed from that of the first insofar as accessibility tests were carried out at a great many points in each canton. (An accessibility test consists in verifying whether a ring tone is obtained when a number is dialled.) The field phase confirmed that the measurement protocol was improved when this test was used.

 

The survey measured coverage along the French road network. Of course, considering the number of cantons surveyed, final conclusions about the whole of France cannot be drawn from the results. However, the results do reveal major differences between cantons, and also between the operators within a given canton. Out of sixty cantons, forty-nine have coverage of more than ninety percent by at least one operator and ten are covered to less than sixty percent by at least one operator. Overall, the three operators provide coverage of eighty-six percent in these cantons. For the one hundred cantons tested during the two coverage surveys carried out by ART this year, average coverage works out at eighty-three percent.

 

ART is making its methodology available to local authorities so that they can carry out surveys to determine, as reliably as possible, the real degree of coverage in the zones that concern them. For this purpose, ART has signed a partnership agreement with an association of départements, and is about to sign several agreements with the départements concerned.

 

The results of these studies will shed light on the choice of coverage objectives for the départements and will make it easier to implement the guidelines laid down during the interministerial committee meeting on regional development, held in Limoges on 9 July 2001.

 

 

RESULTS OF THE SURVEY

Coverage as tested by an investigator standing outdoors and corresponding to calls of acceptable quality

 

Canton

Dpt

Coverage of the operator providing the best coverage in the canton

Coverage of the operator providing the worst coverage in the canton

Accessibility rate (1)

Coverage rate (2)

Accessibility rate (1)

Coverage rate (2)

Saint-Agrève

07

56% ± 1.9%

54% ± 2.3%

38% ± 1.8%

37% ± 2.2%

Pont-Audemer

27

97% ± 0.8%

96% ± 1.4%

84% ± 2.0%

82% ± 2.5%

Authon-du-Perche

28

95% ± 0.8%

93% ± 1.4%

82% ± 1.4%

81% ± 1.9%

Cloyes-sur-le-Loir

28

97% ± 0.4%

96% ± 0.8%

90% ± 0.7%

89% ± 1.1%

Courville-sur-Eure

28

99% ± 0.3%

98% ± 0.7%

91% ± 0.7%

90% ± 1.1%

La Ferté-Vidame

28

97% ± 0.3%

96% ± 0.7%

90% ± 0.6%

87% ± 1.2%

La Loupe

28

98% ± 0.3%

97% ± 0.8%

90% ± 0.7%

89% ± 1.3%

Senonches

28

99% ± 0.4%

98% ± 1.0%

86% ± 1.3%

85% ± 1.7%

Thiron-Gardais

28

99% ± 0.3%

98% ± 0.7%

90% ± 0.9%

88% ± 1.6%

Azay le Rideau

37

91% ± 1.4%

90% ± 1.8%

77% ± 0.2%

75% ± 2.6%

Bourgueuil

37

74% ± 1.5%

73% ± 2.0%

61% ± 1.7%

61% ± 2.0%

Château la Vallière

37

97% ± 0.8%

96% ± 1.4%

55% ± 2.4%

54% ± 2.8%

Chinon

37

97% ± 0.9%

96% ± 1.5%

89% ± 1.7%

87% ± 2.4%

Langeais

37

86% ± 1.9%

84% ± 2.4%

44% ± 2.6%

44% ± 2.9%

L'Ile Bouchard

37

92% ± 1.0%

90% ± 1.5%

82% ± 1.4%

80% ± 1.9%

Neuillé Pont Pierre

37

97% ± 0.9%

97% ± 1.3%

83% ± 1.9%

82% ± 2.4%

Neuvy le Roi

37

94% ± 0.9%

93% ± 1.5%

76% ± 1.7%

74% ± 2.2%

Le Touvet

38

95% ± 0.6%

95% ± 0.9%

89% ± 0.8%

87% ± 1.3%

Lamotte-Beuvron

41

96% ± 0.8%

95% ± 1.4%

81% ± 1.7%

79% ± 2.3%

Meung sur Loire

45

100% ± 0.2%

99% ± 0.7%

96% ± 0.8%

95% ± 1.3%

Figeac est

46

92% ± 0.8%

91% ± 1.2%

77% ± 1.2%

75% ± 1.7%

Duras

47

90% ± 0.9%

89% ± 1.4%

39% ± 1.5%

38% ± 1.9%

La Canourgue

48

84% ± 1.1%

81% ± 1.8%

45% ± 1.4%

43% ± 2.0%

Le Teilleul

50

100% ± 0.2%

99% ± 0.6%

98% ± 0.3%

97% ± 0.9%

Barenton

50

89% ± 1.5%

87% ± 2.1%

79% ± 2.0%

77% ± 2.4%

Ambrières les Vallées

53

100% ± 0.0%

100% ± 0.4%

97% ± 0.5%

95% ± 1.3%

Bais

53

95% ± 1.2%

94% ± 1.7%

49% ± 2.8%

48% ± 3.3%

Couptrain

53

99% ± 0.3%

98% ± 0.9%

92% ± 1.1%

89% ± 1.9%

Gorron

53

99% ± 0.3%

99% ± 0.6%

94% ± 1.1%

92% ± 1.7%

Landivy

53

97% ± 0.8%

96% ± 1.2%

75% ± 1.9%

74% ± 2.4%

Le Horps

53

99% ± 0.5%

98% ± 1.3%

83% ± 2.2%

80% ± 2.9%

Mayenne est

53

100% ± 0.1%

99% ± 0.6%

93% ± 0.6%

90% ± 1.2%

Mayenne Ouest

53

97% ± 0.4%

96% ± 0.9%

87% ± 0.8%

85% ± 1.4%

Villaines la Juhel

53

95% ± 1.0%

94% ± 1.3%

77% ± 1.8%

77% ± 2.1%

Briouze

61

100% ± 0.2%

99% ± 0.7%

71% ± 1.4%

70% ± 1.8%

Carrouges

61

79% ± 2.0%

78% ± 2.4%

69% ± 2.2%

68% ± 2.7%

Courtomer

61

96% ± 0.5%

95% ± 1.0%

84% ± 0.9%

82% ± 1.4%

Ecouché

61

88% ± 1.2%

86% ± 1.8%

53% ± 1.8%

51% ± 2.3%

Exmès

61

99% ± 0.2%

99% ± 0.6%

33% ± 1.2%

31% ± 1.7%

Gacé

61

96% ± 0.7%

95% ± 1.3%

79% ± 1.6%

77% ± 2.2%

Mortrée

61

99% ± 0.3%

98% ± 0.7%

68% ± 1.4%

66% ± 1.9%

Moulins la Marche

61

79% ± 1.1%

77% ± 1.6%

73% ± 1.2%

70% ± 1.8%

Passais

61

98% ± 0.4%

96% ± 0.9%

93% ± 0.7%

92% ± 1.1%

Putanges

61

89% ± 1.1%

88% ± 1.6%

75% ± 1.5%

73% ± 2.2%

Sées

61

97% ± 0.6%

96% ± 1.1%

80% ± 1.4%

78% ± 1.9%

Trun

61

98% ± 0.4%

97% ± 1.0%

56% ± 1.5%

54% ± 2.1%

Tourouvre

61

92% ± 0.9%

91% ± 1.3%

53% ± 1.7%

52% ± 2.1%

Bapaume

62

98% ± 0.4%

97% ± 0.8%

83% ± 1.0%

82% ± 1.5%

Châteldon

63

76% ± 1.3%

74% ± 1.8%

63% ± 1.5%

61% ± 2.0%

Maubourguet

65

99% ± 0.2%

98% ± 0.7%

87% ± 0.8%

85% ± 1.4%

Le Lude

72

97% ± 0.6%

96% ± 1.1%

82% ± 1.3%

80% ± 2.0%

Pontvallain

72

99% ± 0.4%

99% ± 0.8%

78% ± 1.7%

75% ± 2.4%

La Suze sur Sarthe

72

100% ± 0.1%

99% ± 0.6%

99% ± 0.3%

98% ± 0.9%

Loué

72

99% ± 0.3%

99% ± 0.7%

94% ± 0.9%

92% ± 1.4%

Conlie

72

97% ± 0.8%

96% ± 1.4%

84% ± 1.7%

80% ± 2.5%

Sillé le Guillaume

72

94% ± 0.9%

93% ± 1.4%

79% ± 1.5%

78% ± 2.0%

Bourg-St-Maurice

73

98% ± 0.3%

98% ± 0.5%

93% ± 0.5%

92% ± 0.9%

Albi Nord Est

81

99% ± 0.3%

98% ± 0.8%

92% ± 0.7%

91% ± 1.2%

Noirmoutier

81

100% ± 0.1%

100% ± 0.3%

97% ± 0.3%

96% ± 0.6%

St-Georges-les-Baillargeaux

86

99% ± 0.2%

99% ± 0.5%

97% ± 0.4%

96% ± 0.9%

Déplacez le curseur pour consulter le contenu du tableau

(1) Accessibility is defined as the capacity for a user to access the network, that is to obtain the ring tone when making a call, from any point along any road in each of the cantons tested.

 

(2) Coverage is defined as the capacity for a user not only to access the network but also to hold a telephone conversation lasting at least one minute with acceptable sound quality, from any point along any road in each of the cantons tested. This notion of coverage therefore includes the notion of accessibility, with an added criterion in terms of service quality.

The three operators provided average coverage of eighty-six percent in these sixty cantons.

 

Methodology used to assess mobile network coverage

1) General description of the survey

a) Principles

The study consisted in carrying out a field campaign to measure the actual territorial coverage provided by each operator.

The survey tested sixty cantons, fifteen of which were non-contiguous and forty five of which were adjacent.

 

Coverage is defined as the probability that a user can make a call lasting at least one minute, and with satisfactory sound quality, from any point along any road in each of the cantons tested.

 

All measurements were made under conditions reproducing those of a subject outdoors and on foot.

 

The method of statistical measurement used for this study comprised two parts:

  • Measurement: Within the chosen geographic sample area, the following were measured:
  •  

        • Accessibility, by testing whether the ring tone was obtained when the number was dialled
        •  

        • Field strength
        •  

  • Calibration: This part consists in establishing a statistical relationship between the field strength received by the mobile phone and the possibility of actually making a call of acceptable quality and lasting at least one minute, given that the ring tone was obtained.
  •  

 

The rates of coverage provided by each operator were calculated using all geographically representative measurements of field strength and accessibility. The calibration function was used to convert these measurements into a set of probabilities of successfully making a call.

b) Measurement

The inspectors drove around within each canton on all types of road, in all directions and going through all communes, covering an average of 150 km. In each town, the measurement vehicle drove past the town hall.

 

In this way, the investigators were able to cover a substantial proportion – between sixty-five and one hundred percent – of the roads in each canton.

 

Every fifteen seconds without exception, investigators would get out of the vehicle to measure field strength and accessibility.

 

In each canton therefore, the investigators took an average of more than 1,650 field strength and accessibility measurements, that is, more than 550 measurements for each network per canton.

c) Calibration

Depending on the network, between 2,100 and 2,600 calls from a mobile telephone to a wireline were successfully made, that is, the person being called answered the phone.

 

Each call was either cut off or was successfully completed. In this case, sound quality was assessed by two investigators who were talking to one another, one using a cellphone and the other a wireline unit. Three ratings were awarded. A poor rating meant that the investigators had difficulty hearing what the person at the other end was saying. A rating of acceptable meant that the investigators had no difficulty hearing the other person but that the sound quality was poor. A perfect rating meant that sound quality was comparable to that of a wireline. If the investigators' ratings differed, the poorer of the two was used in the test result.

 

In this way analysts were able to establish a calibration law that defined, for each level of field strength, the percentage of successful calls of satisfactory quality that were made, once network accessibility had been obtained, that is, investigators heard the ring tone when they dialled the number.

2) Results and accuracy

The table enclosed shows an assessment of the coverage, that is, the rate of accessibility and the rate of coverage for each canton. The table shows the coverage rate of the operator providing the best coverage in a particular canton and that of the operator with the worst coverage in the same canton. Operator names are not divulged and no information is provided that would allow the operators concerned to be identified. The statistical accuracy of the results is indicated.


Linked documents