Distributed Embedded Systems University of Paderborn From Connected Cars to Smart Ci9es: Novel Applica9ons for Wireless Communica9on Falko Dressler dressler@ccs-labs.org Science Brunch, Zurich From Connected Cars to Smart Ci5es 1
² Mo5va5on Outline ª Trends towards autonomous driving and integrated inter-vehicle communica5on ² Green and safe driving in smart ci5es ª Interac5ons with traffic lights or no traffic lights at all ² Cars as the key ICT component for smart ci5es ª Ubiquitous availability ª Bridging communica5on gaps Science Brunch, Zurich From Connected Cars to Smart Ci5es 2
Towards Autonomous Driving Science Brunch, Zurich From Connected Cars to Smart Ci5es 3
Inter-Vehicle Networking for Situa9on Awareness Illustrations: C2C-CC Science Brunch, Zurich From Connected Cars to Smart Ci5es 4
Towards Heterogeneous Vehicular Networks ² Many communica5on channels available ª Which to pick when? Technology SoXware Defined 2020 Radio? mmwave cellular DSRC VLC Wi-Fi Science Brunch, Zurich From Connected Cars to Smart Ci5es 5
Standardiza9on and Open Research ² Dedicated spectrum for inter-vehicle communica5on in Europe (ECC), the US (FCC), and Japan ² IEEE DSRC/WAVE and ETSI ITS G5 ª Both build upon IEEE 802.11p ² Situa5on awareness as major year-one-applica5on ª ETSI DCC (Decentralized Conges5on Control) ² But many fundamental research ques9ons s9ll unanswered ª Scalability, real-9me capabili9es, use of heterogeneous networks ² Dagstuhl seminar series iden5fied key challenges ª Falko Dressler, Hannes Hartenstein, Onur Al5ntas and Ozan K. Tonguz, "Inter-Vehicle Communica9on - Quo Vadis," IEEE Communica,ons Magazine, vol. 52 (6), pp. 170-177, June 2014. Science Brunch, Zurich From Connected Cars to Smart Ci5es 6
GREEN AND SAFE DRIVING IN SMART CITIES Science Brunch, Zurich From Connected Cars to Smart Ci5es 7
Intersec9on Management ² Efficiency No Traffic Lights Regular Traffic Lights [1] No Traffic Lights Vs Traffic Lights studio tdes CC BY-NC-SA 2.5 hnp://www.youtube.com/watch?v=hfoo3e0nxsi Science Brunch, Zurich From Connected Cars to Smart Ci5es 8
Intersec9on Management ² Safety No Traffic Lights [1] Авария! Проигнорировал знак стоп SmexaMnogo, CC BY hnp://www.youtube.com/watch?v=to-ernw1hto Science Brunch, Zurich From Connected Cars to Smart Ci5es 9
Safe and Efficient Traffic in Smart Ci9es ² STEP 1: Communica5on between traffic lights and vehicles ² First ideas date back to the 1970ies ª ª ² Reality today especially in Asian countries ª ² Analog communica5on Transmission of switching 5mes to approaching vehicles Large displays indica5ng the end of the current light phase Vehicular networking enables the solu5ons for tomorrow ª ª WiFi or DSRC based Inves5gated, for example, in Travolu5on project Science Brunch, Zurich From Connected Cars to Smart Ci5es 10
Safe and Efficient Traffic in Smart Ci9es ² STEP 2: Enabling Virtual Traffic Lights ² Status as of today ª Only 1-20% of all intersec5ons are equipped with traffic lights ª Many of these are not appropriately connected à may cause ar5ficial traffic jams ª New installa5ons are very expensive ² VTL Virtual Traffic Lights ª First inves5gated by CMU/U Porto Science Brunch, Zurich From Connected Cars to Smart Ci5es 11
Virtual Traffic Lights 1. Wait for intersec5on 2. Perform leader elec5on 3. Send/receive TL program ² Research avenues: ª Leader elec5on leader ª TL program computa5on ª Scalability and load? Science Brunch, Zurich From Connected Cars to Smart Ci5es 12
Virtual Traffic Lights Science Brunch, Zurich From Connected Cars to Smart Ci5es 13
Safe and Efficient Traffic in Smart Ci9es ² STEP 3: Using Heterogeneous Networks ² Today ª Bluetooth ª WiFi (IEEE 802.11a/b/g/n/ac) ª 3G/4G (UMTS/LTE) ² Tomorrow ª DSRC/WAVE (IEEE 802.11p) ª 60 GHz (IEEE 802.11ad) ª 4G+ (LTE Advanced/LTE Direct) ª Millimeter wave (Radar) ª Visible light communica5on Science Brunch, Zurich From Connected Cars to Smart Ci5es 14
Safe and Efficient Traffic in Smart Ci9es ² STEP 3: Using Heterogeneous Networks ² Causes new challenges but provides many new opportuni5es ª Clustering ª Selec5ng the communica5on channel ª Coopera5on and self-organiza5on Science Brunch, Zurich From Connected Cars to Smart Ci5es 15
VEHICLES AS THE INFORMATION HUB IN SMART CITIES Science Brunch, Zurich From Connected Cars to Smart Ci5es 16
Self-Organizing Network of Cars ² ² can serve as distributed data store and retrieval mechanism May include even parked cars as informa5on highway how does when will she highway 1 look get here? right now? where can I park my car? is this the end how is traffic of the jam? on 2nd street? Science Brunch, Zurich From Connected Cars to Smart Ci5es 17
Making Cars the Main ICT Component 1. Sharing data with other users. 2. Sharing data with specific loca5on. 3. Uploading data without direct Internet access. 4. Live data from disaster area. 5. Wide Area Data Storage. 6. Wide Area Distributed Compu5ng. 7. Disaster Area V2X Evacuation Site V2V Internet Drive with messages Science Brunch, Zurich From Connected Cars to Smart Ci5es 18
Making Cars the Central ICT Component ² Cars are ubiquitously available ² Feature all kinds of communica5on technology ² Relaying of messages ª has been shown to work quite well ª but will fail in low density/penetra5on areas ² Parked Cars ª are readily available, especially in low-density areas ª park in strategically promising posi5ons along the street and next to obstacles ª can be used as relaying nodes in order to increase coopera5ve ª awareness for moving vehicles Science Brunch, Zurich From Connected Cars to Smart Ci5es 19
Scenario? Science Brunch, Zurich From Connected Cars to Smart Ci5es 20
Besides communica9on vehicles provide Car Sensor Battery Communication Processor Storage Use case Use case Use case Use case Use case Use case Use case Use case Use case Use case Vehicular resources are available not only for vehicles themselves but also for humans and things Vehicles store, process, and convey the information generated by human or things. Science Brunch, Zurich From Connected Cars to Smart Ci5es 21
Naming ² Using concepts from NDN/ICN ª Hash tags to iden5fy services ª Meta data for fine grained descrip5on Hash tag Meta data User hash(file1) location = Tokyo, type = video, size = 2GB 1 hash(file1) location = Tokyo, type = video, size = 2GB 3 hash(file1) location = Paderborn, type = video, size = 2GB 7 hash(file2) location = Tokyo, type = image 1 hash(file2) type = image, size = 500MB 12 CPU location = Paderborn, type = ARM 7 Storage location = Paderborn, type = hours, size = 78GB 7 * type = image hash(file1) * 1, 12 1, 3, 7 Science Brunch, Zurich From Connected Cars to Smart Ci5es 22
Discovery Latency in an Urban Environment Science Brunch, Zurich From Connected Cars to Smart Ci5es 23
Conclusions Today, we studied ² Challenges and opportuni5es of using connected cars concepts ª Capability to connect everyone and everything ª Can be seen as a big data storage ª Help improving our daily road traffic experience and safety ² Not discussed ª Security issues: Strong debate about privacy vs. security as can be seen, there are many open challenges and ques9ons for another decade of interes9ng research J Science Brunch, Zurich From Connected Cars to Smart Ci5es 24
More Informa9on? Vehicular Networking (Cambridge University Press) IEEE Vehicular Networking Conference (VNC) ² Tokyo, New Jersey, Amsterdam, Seoul, Boston, Paderborn 20% off ² Next: Kyoto in Dec 2015 Science Brunch, Zurich From Connected Cars to Smart Ci5es 25