Giornata di studio sui White Spaces, che si è tenuta lunedì 24 ottobre 2011, 9.30 - 17.00 – Sala Agorà, Incubatore d’Imprese Innovative del Politecnico di Torino (I3P).
Time Series Foundation Models - current state and future directions
La Sperimentazione White Spaces in Piemonte Applicazioni per Reti Veicolari
1. La Sperimentazione White Spaces
in Piemonte
Applicazioni per Reti Veicolari
Claudio Casetti
Dipartimento di Elettronica
Politecnico di Torino
2. Goals
¤ ITS (Intelligent Transportation Systems) environment
¤ Design a roadside downloading application for vehicles
with prefetching of content
¤ Exploit long-range UHF coverage for signalling traffic
(vehicle position, content requests)
¤ Exploit short-range, high-bandwidth 5GHz links for content
downloading
3. Starting Point
¤ VICSUM project
¤ funded by Regione Piemonte in 2007-09
¤ partners: Politecnico di Torino, CSP, CRF
¤ The VICSUM legacy
¤ Hand-on expertise in VANETs
¤ Designing mobile applications
¤ Living testbeds (Villa Gualino)
¤ The “VICSUM cars”
4. Network Architecture
¤ Infrastructure
¤ Central Controller (CC)
¤ Roadside Units (RSU)
¤ Long-Range Unit (LRU)
¤ Vehicle On-Board Units (OBU)
RSU1 RSU2
5. Network Architecture
¤ Control Plane
¤ Common channel shared by CC (though LRU) and Vehicle
OBUs
(700 MHz)
¤ High-bit rate communication on point-to-point link between
CC and RSU
¤ User Plane
¤ RSU to vehicles OBU communication (5 GHz)
6. The LFT Protocol
¤ We designed and implemented a communication
protocol called Locate-Fetch-Transfer (LFT)
¤ LFT operates among RSUs, CC and OBUs
¤ LFT incorporates control and user plane
7. LFT Protocol Overview (1)
¤ Vehicle notifies to CC over
700 MHz link:
¤ own position
¤ content requests
¤ CC fetches requested content
and dispatches it to
the nearest RSU
RSU1 RSU2
¤ Vehicle associates to RSU and
downloads as much as
possible
8. LFT Protocol Overview (2)
¤ When moving out of RSU
coverage, the download stops
¤ RSU reports the outcome of
the download
¤ If incomplete, the CC
schedules the download to
continue at the next RSU
RSU1 RSU2
¤ Remaining file is dispatched
to the next RSU
9. The Test
¤ The OBU requests the download of a 200-MB file
¤ The download is completed over several “contacts”
with RSUs
¤ We measured
¤ the number of contacts to complete the transfer
¤ the coverage time
¤ the achieved throughput
10. The Test
¤ Preliminary tests at Villa Gualino (Torino)
¤ Signalling traffic on a 2.4-GHz (802.11b) channel
¤ Data channel on a 5-GHz (802.11a) channel
¤ Final tests in Val di Viù
12. Results
20 km/h 40 km/h
25 25
vehicle request and vehicle configuration vehicle request and vehicle configuration
vehicle configuration vehicle configuration
20 20
Throughput [Mbps]
Throughput [Mbps]
15 15
10 10
5 5
0 0
0 100 200 300 400 500 600 700 0 200 400 600 800 1000 1200
Time [s] Time [s]
13. Comparison with simple transfer
Our protocol: 7 contacts needed to transfer content
Access Point AP1 AP2 AP2 AP1 AP1 AP2 AP2
# of downloaded chunks
from the CC 35000 35000 35000 35000 35000 35000 18494
# of downloaded chunks 32850 30430 24750 18686 12650 28950 18494
Coverage time [s] 39 32 41 48 33 31 20
Throughput [Mb/s] 8.33 9.4 5.97 3.85 3.79 9.23 9.14
Simple transfer: 10 contacts needed to transfer content
Access Point AP1 AP2 AP2 AP1 AP1 AP2 AP2 AP1 AP1 AP2
# of downloaded chunks
from the CC 35000 35000 35000 35000 35000 35000 35000 35000 35000 17379
# of downloaded chunks 17890 22887 26080 21261 29190 19903 32574 27438 26555 17379
Coverage time [s] 38 40 48 46 35 36 40 38 39 21
Throughput [Mb/s] 3.68 3.77 5.37 3.75 6.41 3.45 7.81 5.8 5.39 4.52
14. Conclusions
¤ LFT is not a Cognitive Radio application but could work
with any CR system
¤ Sustained throughput comparable to ADSL speeds to
mobile users
¤ Use of today’s technology
¤ Future work: lots of optimization to do!