Date post: | 28-Mar-2015 |
Category: |
Documents |
Upload: | emerson-marland |
View: | 217 times |
Download: | 3 times |
Copyright © Siemens plc 2009. All rights reserved.
Above ground SCOOT and MOVA detection in practice
Page 2September-09 Siemens Mobility / Traffic Solutions
Backgound
SCOOT and MOVA are the strategies of choice for many Traffic Managers
Both require accurate data about traffic conditions
Traditionally this is supplied by loops
Loops have many advantages, but also some disadvantages
Obvious solution is to use above ground solutions!
Advantages of loops
• Accurate• Precise detection zone and position• Immune to weather and lighting conditions• Almost invisible when installed• Can’t easily be stolen
Disadvantages of loops
• Difficult to install• Vulnerable to damage
• Movement of the road surface• Destruction when services are accessed
• Difficult to repair
Page 3September-09 Siemens Mobility / Traffic Solutions
Possible above ground technologies
Several alternatives sensor technologies assessed by Siemens Corporate Technology (CT)
Some found more suitable than others
77GHz radar seemed the best option
Advantages
• Passive so can be very low power• Accurate zones if multiple sensors used• Immune to changing light levels• Relatively low cost• No significant maintence required
Disadvantages
• Not good for detecting slow / static traffic
• Can be affected by environment (rain / snow etc)
Passive Acoustic
Technology
• Immune to light levels• Relatively low cost• No significant maintenance required
• Difficult to achieve precise detection zone
• Heavy rain / wind and snow can severally affect performance
Active Acoustic
• Immune to changing light levels• Very accurate zones possible• No significant maintenance required
• Active device so needs radio approval in all target markets
77GHz Radar
• Passive so can be very low power• Reasonably accurate zones possible• Relatively low cost• No significant maintenance required
• Can be affected by significant sudden temperature shifts
• Fog and heavy snow can affect performance
Passive Infrared
• Many accurate zones possible (if field of view permits)
• Can provide visual overview of target area
• Can be badly affected by changing light levels and environmental conditions (Rain, snow etc).
• Significant maintenance overhead
Video
Page 4September-09 Siemens Mobility / Traffic Solutions
Benefits of 77GHz radar
Excellent target resolution Able to deliver very precise
detection zones
Important if SCOOT and MOVA loops are to be emulated accurately.
Good immunity to adverse weather
Small antenna size Minimal visual impact on-street
Widespread acceptance of the frequency band
Detector could be used worldwide with few limitations
Page 5September-09 Siemens Mobility / Traffic Solutions
Benefits of 77GHz radar
But there was a big problem
Cost!!!
Detailed evaluation showed although it offered a very good technical solution, solid state 77GHz technology was in its infancy and the product costs were found to be very high.
However, radar operating at 24GHz seemed to offer similar benefits at a much lower cost
Page 6September-09 Siemens Mobility / Traffic Solutions
Development challenges
Product cost is a very significant driver when choosing technology solutions
If these are too high product adoption in the market will tend to be limited, even if the performance achieved is good
One interesting solution relates to the cost optimisation of the radar antenna and related circuits
Page 7September-09 Siemens Mobility / Traffic Solutions
Experiences in practice
First tests based on “ground truth” principals at Poole test site
Loop comparison Various traffic and weather
conditions
Initially only count and occupancy gathered in 5 minute bins
Results looked very good
Page 8September-09 Siemens Mobility / Traffic Solutions
Experiences in practice
Flow : Loop Versus Heimdall (20/5/09 to 29/5/09)
0
100
200
300
400
500
600
700
800
900
1000
00:0
0
06:0
0
12:0
0
18:0
0
00:0
0
06:0
0
12:0
0
18:0
0
00:0
0
06:0
0
12:0
0
18:0
0
00:0
0
06:0
0
12:0
0
18:0
0
00:0
0
06:0
0
12:0
0
18:0
0
00:0
0
06:0
0
12:0
0
18:0
0
00:0
0
06:0
0
12:0
0
18:0
0
00:0
0
06:0
0
12:0
0
18:0
0
00:0
0
06:0
0
12:0
0
18:0
0
00:0
0
06:0
0
Time
Veh
icle
Co
un
t
Loop
Heimdall
Reference loop v Heimdall
Page 9September-09 Siemens Mobility / Traffic Solutions
Experiences in practice
First tests based on a “ground truth” approach at Poole test site
Loop comparison Various traffic and weather
conditions
Initially only count and occupancy gathered in 5 minute bins
Results looked good
But its is easy to be fooled by simple graphs!
Focusing on the variation between the loop and Heimdall seemed to showed a different picture
Page 10September-09 Siemens Mobility / Traffic Solutions
Experiences in practice
Delta Flow (Percentage Variation) Between Loop and Heimdall
-25
-15
-5
5
15
25
00
:00
11
:00
22
:00
09
:00
20
:00
07
:00
18
:00
05
:00
16
:00
03
:00
14
:00
01
:00
12
:00
23
:00
10
:00
21
:00
08
:00
19
:00
06
:00
17
:00
04
:00
Time
% V
ari
ati
on
Delta (%)
Difference between loop v Heimdall (Extract)
Page 11September-09 Siemens Mobility / Traffic Solutions
Delta Flow (Percentage Variation) Between Loop and Heimdall
-25
-15
-5
5
15
25
00:00
11:00
22:00
09:00
20:00
07:00
18:00
05:00
16:00
03:00
14:00
01:00
12:00
23:00
10:00
21:00
08:00
19:00
06:00
17:00
04:00
Time
% Va
riatio
n
Delta (%)
Difference between loop v Heimdall (Extract)Delta Flow (Percentage Variation) Between Loop and
Heimdall
-25
-15
-5
5
15
25
00:00
11:00
22:00
09:00
20:00
07:00
18:00
05:00
16:00
03:00
14:00
01:00
12:00
23:00
10:00
21:00
08:00
19:00
06:00
17:00
04:00
Time
% Va
riatio
n
Delta (%)
Difference between loop v Heimdall (Extract)
Experiences in practice
Biggest variation occurred at night
few vehicle in the sample set
loop was found to be missing some fast moving motor cycles which Heimdall detected.
Also the loop detected a number of vehicles driving close to the loop but in the opposite direction,
Its easy to adjust Heimdall to avoid these vehicles
Overall conclusion was that Heimdall count and occupancy performance looks good
Page 12September-09 Siemens Mobility / Traffic Solutions
Implications SCOOT
SCOOT uses a measure of linear discounted occupancy to model queue build-up and dispersal (Link Profile Units - LPUs)
The way a detector responds can significantly affect the LPU values that SCOOT calculates
Typical detector sample
250ms data 0 0 1 1 1 1 0 0 0 0 0 1 1 1 1 1 0 0 0 0 1 1 1 0 0 1 1 0 0 0Discounted occupancy 7 6 5 4 7 6 5 4 3 7 6 5 7 6Resulting LPU value = 78 22 1825 13
250ms data 0 0 0 1 1 1 0 0 0 0 0 0 1 1 1 1 0 0 0 0 0 1 1 0 0 0 1 0 0 0Discounted occupancy 7 6 5 7 6 5 4 7 6 7Resulting LPU value = 60 718 22 13
Detector with poor turn-on response
250ms data 0 0 1 1 0 1 0 0 0 0 0 1 1 1 0 1 0 0 0 0 1 0 1 0 0 1 1 0 0 0Discounted occupancy 7 6 7 7 6 5 7 7 7 7 6Resulting LPU value = 72 7 7 7 713 18 13
Detector with poor hold characteristics
Count value accurate
LPU value significantly lower
Count value high
LPU value only slightly lower (but likely not be consistent)
Page 13September-09 Siemens Mobility / Traffic Solutions
Implications SCOOT – LPU on test site
Initial Heimdall LPU performance was disappointing!
Further refinement of the detection algorithm proved necessary
Performance much improved as a result.
Still produces a slight lower LPU value than the loop
Heimdall detection zone about 10% smaller than the loop
Would be accounted for during normal validation of the link
LPU : Loop Versus Heimdall (20/5/09 to 29/5/09)
0
50
100
150
200
250
300
350
00:0
0
06:4
5
13:3
0
20:1
5
03:0
0
09:4
5
16:3
0
23:1
5
06:0
0
12:4
5
19:3
0
02:1
5
09:0
0
15:4
5
22:3
0
05:1
5
12:0
0
18:4
5
01:3
0
09:1
5
16:0
0
22:4
5
05:3
0
12:1
5
19:0
0
01:4
5
08:3
0
15:1
5
22:0
0
04:4
5
11:3
0
18:1
5
01:0
0
07:4
5
Time
LP
U
Loop
Heimdall
LPU:- Loop v Heimdall
Page 14September-09 Siemens Mobility / Traffic Solutions
Heimdall extended trials
Several further Heimdall trial installations have been undertaken
Different arrangements have been tested,
Live installations With and without reference
loops
A full set of detailed results on all the trials will be available from the Siemens website
Poole – Fleetsbridge Roundabout (SCOOT – count site)
Southampton – Bittern Triangle (SCOOT)
Bournemouth – Cemetery Junction (SCOOT – count site)
Cardiff - Leckwith Giratory (SCOOT)
Winchester. Battery Hill (SCOOT)
Wimborne - Willett Arms (MOVA)
Heimdall trial sites
Page 15September-09 Siemens Mobility / Traffic Solutions
Winchester – Battery Hill
A single SCOOT detector has been installed in Winchester, along side a reference loop used by TRL.
Data still being gathered from the Hampshire UTC system.
Count correlation between loop and Heimdall were very good.
LPU results re-confirmed that the Heimdall detection zone is a little smaller than a standard loop, but acceptable
A view confirmed by a TRL assessment of the data from this site
Poole – Fleetsbridge Roundabout (SCOOT – count site)
Southampton – Bittern Triangle (SCOOT)
Bournemouth – Cemetery Junction (SCOOT – count site)
Cardiff - Leckwith Giratory (SCOOT)
Winchester. Battery Hill (SCOOT)
Wimborne - Willett Arms (MOVA)
Heimdall trial sites
Romse
y Roa
d
Battery Hill
SCOOT Loop versus Heimdall - FlowNov. 1st - Nov. 10th 2008
0
100
200
300
400
500
600
700
800
900
00:00
05:15
10:30
15:45
21:00
02:15
07:30
12:45
18:00
23:15
04:30
09:45
15:00
20:15
01:30
06:45
12:00
17:15
22:30
03:45
09:00
14:15
19:30
00:45
06:00
11:15
16:30
21:45
03:00
08:15
13:30
18:45
00:00
05:15
10:30
15:45
21:00
02:15
07:30
12:45
18:00
23:15
04:30
09:45
15:00
20:15
Time
Flow
(Veh
icle C
ount
)
loop RFLOW
heim RFLOW
Battery Road – LPU (Reference loop v Heimdall)
LPU
50
100
150
200
250
300
350
400
450
LoopHeimdall
SCOOT Loop versus Heimdall - FlowNov. 1st - Nov. 10th 2008
0
100
200
300
400
500
600
700
800
900
00:00
05:15
10:30
15:45
21:00
02:15
07:30
12:45
18:00
23:15
04:30
09:45
15:00
20:15
01:30
06:45
12:00
17:15
22:30
03:45
09:00
14:15
19:30
00:45
06:00
11:15
16:30
21:45
03:00
08:15
13:30
18:45
00:00
05:15
10:30
15:45
21:00
02:15
07:30
12:45
18:00
23:15
04:30
09:45
15:00
20:15
Time
Flow
(Veh
icle C
ount
)
loop RFLOW
heim RFLOW
Battery Road – LPU (Reference loop v Heimdall)
LPU
50
100
150
200
250
300
350
400
450
LoopHeimdall
Page 16September-09 Siemens Mobility / Traffic Solutions
Summary
The Heimdall detector provides good correlation to a loop
Count accuracy is excellent
SCOOT LPU values are a little lower than a standard loop but this can be addressed during validation
MOVA performance has also been shown to be acceptable
A full report on the Heimdall trials will be available from the Siemens website
Copyright © Siemens plc 2009. All rights reserved.
Above ground SCOOT and MOVA detection in practice