Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
Should humans fly in aerial vehicles?
In the majority of urban air mobility planning, goods such as Amazon packages and food are the targeted cargo population that urban air traffic will predominantly transport. Furthermore, this traffic will likely be conducted autonomously or in a semi-supervised style. This is in stark contrast to ground transportation, in which human-operated vehicles are the dominant population of ground
transportation users.
However, humans may wish to fly and/or navigate aerial vehicles themselves — for legal liability, for
improvements in safety/efficiency, and/or for recreational purposes. While human-operated vehicles are
most definitely less efficient than autonomous UAVs, it is not immediately obvious that the efficiency of
overall air traffic will decrease when introducing human-operated aerial vehicles. Will air traffic efficiency
decrease when human-navigated aerial vehicles operate alongside autonomous aerial vehicles in urban
skies, and are there situations in which introducing human traffic will improve air traffic efficiency?
Mehr [1] introduced a two-population routing game model to analyze the effectiveness of human-autonomous mixed traffic. The sensitivity of social welfare concerning the human-
autonomous population ratio, as well as the traffic network structure, is analyzed. In particular, Mehr
showed that when the ratio of the load capacities between two populations is homogeneous, the total
delay in the network will not increase as the ratio of autonomous vehicles increases. On the contrary,
when these conditions are not satisfied, there exist many networks in which the total delay will increase
as the presence of autonomous vehicles increases.
To answer the question posed in the project title, we will apply the same analysis to a dynamic,
stochastic version of routing games: Markov decision process (MDP) congestion games [2]. We will
focus on developing 1) an accurate model of human traffic vs autonomous traffic and 2) a mathematical
characterization of routing networks in the aerial setting.
Mehr [1] introduced a two-population routing game model to analyze the effectiveness of human-autonomous mixed traffic. The sensitivity of social welfare concerning the human- autonomous population ratio, as well as the traffic network structure, is analyzed. In particular, Mehr showed that when the ratio of the load capacities between two populations is homogeneous, the total delay in the network will not increase as the ratio of autonomous vehicles increases. On the contrary, when these conditions are not satisfied, there exist many networks in which the total delay will increase as the presence of autonomous vehicles increases.
To answer the question posed in the project title, we will apply the same analysis to a dynamic, stochastic version of routing games: Markov decision process (MDP) congestion games [2]. We will focus on developing 1) an accurate model of human traffic vs autonomous traffic and 2) a mathematical characterization of routing networks in the aerial setting.
1. Understand routing games, MDP congestion games, braess paradox
2. Understand the Nash equilibrium conditions and its KKT characterization in the potential game setting
3. Understand the impact of aerial networks on the Nash equilibrium and social welfare
1. Understand routing games, MDP congestion games, braess paradox 2. Understand the Nash equilibrium conditions and its KKT characterization in the potential game setting 3. Understand the impact of aerial networks on the Nash equilibrium and social welfare
Please send a short description of why you are interested, your resume/CV (including lists of relevant publications/projects) and a transcript of records in PDF format via email to
sarahli@control.ee.ethz.ch and negar@berkeley.edu
Please send a short description of why you are interested, your resume/CV (including lists of relevant publications/projects) and a transcript of records in PDF format via email to sarahli@control.ee.ethz.ch and negar@berkeley.edu
SAMA_mixed_ground_air_mobility.pdf
