Emergency Braking: A Study of Network and ... - Semantic Scholar

0 downloads 127 Views 517KB Size Report
Nov 30, 2011 - Emergency Braking: A Study of Network and Application ... Brake Lights (EEBL) ... a probability P which d
Emergency Braking: A Study of Network and Application Performance --Michele Segata, Renato Lo Cigno, VANET, 2011.

Renyong Wu 2011.11.30

Motivation • The interaction of the communications (protocols) with the car’s adaptive cruise control (ACC) system and the driver’s behavior for cars that are not equipped with these communication devices. • Considering Collisions and combining driver’s behavior instead of crash-free • Just a simulation study (ns-3) on Emergency Electronic Brake Lights (EEBL)

1. Introduction Questions: • How do EEBL and the VANET network actually interact? • What is the market penetration rate that makes EEBL benefits measurable? • Do non-equipped vehicles benefit from a partial EEBL introduction? • What is the most effective communication range? • Are rebroadcast techniques needed to make EEBL more effective? If yes, how do they impact the network load? • Can we find easy and efficient message aggregation techniques?

3. Simulation Models and Tools Introduce a maximum deceleration (10m/s) as a physical limit in addition to the value obtained from the IDM formula:

3. Simulation Models and Tools The safe gap depends on the speed of the current vehicle and is computed:

If actual gap is lower than the safe gap, the follower brakes; else computes acceleration

3. Simulation Models and Tools • Broadcast frequency: 10Hz, which means 100ms. Here plain EEBL means without rebroadcast. • P-persistence: The rebroadcast decision is taken with a probability P which depends on the distance from the sender.  EEBLR • Every 100ms, the queue is empted and a single frame aggregating all messages is sent.  EEBLA

4. Network Level Results Two metrics to measure the network performance: • Channel load at each station ρi(t); • The percentage of frames luf that are not received by any station.

4. Network Level Results

4. Network Level Results

A color plot of the load ρi(t) as a function of time and space. Note around the time instant 130s the braking begins.

5. Application Performance

Percentage of cars involved in accidents vs. Market Penetration Rate for single-lane tests for different protocols and different average speeds

5. Application Performance

Percentage of cars involved in accidents vs. Market Penetration Rate for multi-lane tests, average speed 130 km/h, two and five-lane scenarios

5. Application Performance

6. Conclusions and Future Work • The time-space analysis of the channel load is novel, as it gives insight in the dynamics of the joint network and application evolution .