Student projects

Low-cost air quality sensors are increasingly deployed in cities to complement sparse reference-grade monitoring stations. However, their measurements are affected by drift, environmental sensitivity, and manufacturing variability, which limits their direct use for reliable air quality assessment. Calibration is therefore essential to improve their accuracy and usability. This project aims to systematically explore and compare calibration techniques for low-cost air quality sensors, using open datasets that include both reference and low-cost measurements. The primary goal is to assess which calibration strategies are most effective under different urban conditions and pollutant types.

During this project, you are expected to:

• Collect and preprocess datasets containing co-located measurements from low-cost sensors and reference stations.
• Implement and test different calibration techniques, such as linear regression, polynomial mapping, Gaussian process regression, and machine learning methods (e.g., random forest, neural networks).
• Evaluate calibration performance across pollutants (e.g., PM2.5, NO₂) and conditions (e.g., seasons, temperature ranges).
• Summarize findings into practical recommendations, highlighting trade-offs between accuracy, robustness, and computational cost.

Contact: Anna Paulish ([email protected])

Open air quality datasets released by cities worldwide offer valuable opportunities to understand urban pollution dynamics, but they vary widely in terms of sensor density, spatial coverage, and temporal consistency. This project aims to perform a systematic exploratory analysis of such datasets, focusing on characterizing coverage, variability, and potential sources of uncertainty. The primary goal is to benchmark different datasets and provide practical insights into their suitability for spatio-temporal modeling.

During this project, you are expected to:

• Collect and preprocess one or more open urban air quality datasets (e.g., PM2.5, NO₂).
• Analyze and visualize spatial coverage, temporal resolution, and data completeness.
• Identify and quantify noise patterns, missing data, and inconsistencies.
• Provide summary statistics and visual reports that can support future modeling efforts.

Contact: Anna Paulish ([email protected])

Electric ride-hailing vehicles are increasingly prevalent. Because these vehicles rely heavily on public charging infrastructure, their charging behaviors impose substantial challenges on the power system. Current research often makes assumptions that electric ride-hailing vehicles get fully recharged when leaving the charging stations. In practice, however, drivers may partially recharge to reduce downtime or respond to nearby demand. Analyzing these behaviors with real-world data enables a more accurate understanding of charging demand and better informs infrastructure planning. This project will examine the partial charging behavior using taxi and charging-station data from Shenzhen.

During this project, you are expected to 

• Process historical Shenzhen electric taxi trajectory and charging station data and identify charging behaviors.
• Identify partial charging behaviors based on consecutive charging and operation sessions of individual electric taxis. 
• Analyze the spatiotemporal patterns of partial charging behaviors.

Contact: Xuhang Liu ([email protected])

Mobility-as-a-Service (MaaS) aims to provide seamless multi-modal mobility options to travelers and has great potential to improve the accessibility of urban transportation systems. With the increasing availability of convenient transfers and easily accessible travel option information, travelers are now faced with a greatly enlarged set of possible multi-modal trips. Understanding travelers’ preferences for these trips, as well as the factors shaping their choices, is therefore essential. Such insights can guide the design of MaaS platforms, optimize service offerings, and enhance user satisfaction by aligning system design with actual traveler behavior.

Building on a previous student project that developed a multimodal travel option searching algorithm and data of collected multi-day GPS trajectories of individuals in Switzerland, this project seeks to investigate the primary factors influencing travelers’ mode combinations and route choices, providing a data-driven foundation for improving multi-modal mobility services.

During this project, you are expected to 

• Review the literature of discrete choice models for mode and route choice behavior.
• Detect multi-modal trips and mode changes during the trips using multi-day GPS trajectory data.
• Explore key factors influencing mode combinations and route choices in multimodal trips.

Contact: Xinyu Ma ([email protected])

On-demand meal delivery service (e.g., UberEat) has developed rapidly in recent years, particularly during COVID lockdowns. Similar to ride-hailing platforms, on-demand meal delivery platforms promptly collect orders and assign them to couriers for pickup and delivery. Differently, some platforms take the bundling strategy that groups several orders with close pickup and delivery locations into bundles. Accordingly, each courier service trip consists of multiple pickups and deliveries. Typically, solving the order bundling requires extensive evaluations on candidate bundles and each evaluation involves solving a Pickup-and-Delivery-Route-Planning (PDRP) problem. This imposes great computational challenge for real-time operations. This project aims to develop a reinforcement learning framework to tackle the order bundling problem. 

During this project, you are expected to 

• Learn the agent-based on-demand meal delivery simulation framework
• Formulate the order bundling problem in the reinforcement learning framework
• Design state variables, reward function, and value/policy network structure
• Perform model training, testing, and parameter tuning
• Evaluate the performance compare them with benchmark algorithms

Contact: Prof. Kenan Zhang ([email protected]) 

Mobility-as-a-Service (MaaS) is an emerging concept that integrates transport services from both public and private transport service providers (TSPs) through a unified platform, enabling users to plan, book, and pay for multi-modal trips via a single account. A common business model of MaaS is to operate as an intermediary: the platform purchases service capacities from TSPs, then designs and sells multi-modal trips to travelers. As a new market player, a MaaS platform must carefully design pricing schemes that not only attract users but also ensure the participation and cooperation of TSPs. This relationship is often referred to as coopetition.

This project will develop a simulation-based model of day-to-day dynamics, in which the MaaS platform and TNCs iteratively revise their pricing and capacity strategies in pursuit of profit maximization, while travelers adapt their service choices in response to evolving trip prices and service levels (e.g., pickup waiting times determined by available capacity). The project aims to reveal the evolutionary dynamics of coopetition between the MaaS platform and TNCs, focusing on how their strategies interact and evolve over time. By capturing the feedback loop between supply-side strategies and demand-side behavior, the study will provide insights into the implementation and long-term profitability of MaaS platforms in competitive mobility markets.

Supervisor: Xinyu Ma

Student: Mathis Magnin

Network design problem (NDP) refers to the optimization of link (node) location and attribute (e.g., capacity) subject to both physical and economic constraints. Classic NDPs often employ system-level aggregate metrics as their objectives, such as minimizing total travel time or maximizing total travel utility, whereas ignoring the variation in individual welfare and thus possibly resulting in discrimination on some travelers. To address this issue, this project will explore accessibility-based NDP that aims to provide sufficient accessibility for all travelers to reach most destinations within a certain time budget.  

Supervisor: Prof. Kenan Zhang

Student: Henri Collet

Commercial speed, defined as the average travel speed of public transport vehicles from one stop to another, is a crucial indicator of the efficiency and robustness (“health”) of public transit networks. Due to traffic congestion, lack of coordination with traffic signals, temporal high passenger volume, complex interactions with other vehicles on roads, and other unforeseen events, real-world operations of public transit often deviate from designed schedules and experience significant delays. 

In collaboration with TPG, the major transit operator in Geneva, this project aims to perform a data-driven diagnostics on the current state of public transit networks. We will leverage a large dataset that collects individual vehicles over the past ten years and focus our analysis on the commercial speed with a primary goal to identify where and why operational bottlenecks occur repeatedly. 

Supervisor: Dr. Cloe Cortes (TPG), Prof. Kenan Zhang

Students: Luca Liuzzi, Ching-Chi Chou

Traffic congestion has long been a critical issue in large cities. As the well-known solution, congestion pricing is however arguably unfair as it tends to favor wealthier travelers. To address such an equity issue, we proposed a mechanism named CARMA that makes travelers to bid for access to scarce transport resources with non-tradable mobility credits, in anticipation of other travelers’ decisions as well as their own future trips. In our preliminary study, we numerically showed the redistribution of credits plays a key role in the system efficiency and found its strong connection to classic road pricing schemes. This project aims to further investigate credit redistribution as a design problem and deepen the understanding of its impact on system performance. 

Supervisor: Prof. Kenan Zhang

Student: Mathis Magnin

Order bundling in on-demand meal delivery service (e.g., UberEat) requires extensive evaluations on candidate bundles and each evaluation involves solving a Pickup-and-Delivery-Route-Planning (PDRP) problem. This imposes great computational challenge for real-time operations. In this project, we analyzed a real-world dataset of meal delivery service and explored the key factors affecting the order bundling decisions. 

Supervisor: Xuhang Liu

Student: Margot Chapalain, Malak Jria

Order bundling in on-demand meal delivery service (e.g., UberEat) requires extensive evaluations on candidate bundles and each evaluation involves solving a Pickup-and-Delivery-Route-Planning (PDRP) problem. This imposes great computational challenge for real-time operations. In this project, we analyzed a real-world dataset of meal delivery service and explored the key factors affecting the order bundling decisions. 

Supervisor: Anke Ye, Kenan Zhang

Student: Elias Rafoul

Mobility-as-a-service (MaaS) aims to provide seamless multi-modal mobility options to traveler and has great potentials to improve sustainability and accessibility of urban transportation systems. One critical factor in MaaS system design, however often overlooked, is the transfer cost, e.g., extra walk and wait due to mode transfers. Using the multi-day GPS trajectories of individuals collected in Switzerland, this project explored how travelers move through the multi-modal transport system, identify different segments of their trips, and detect mode changes and transfers. It contributes to a better understanding of real-life travel behaviors and provides useful insights to design MaaS systems.

Superviser: Rui Yao, Xinyu Ma

Students: Hana Wermeille, Orange Koenga

Mobility-as-a-service (MaaS) aims to provide seamless multi-modal mobility options to traveler and has great potentials to improve sustainability and accessibility of urban transportation systems. One critical factor in MaaS system design, however often overlooked, is the transfer cost, e.g., extra walk and wait due to mode transfers. With the operational and planning data provided by all operators in Switzerland, this project constructs a time-specific multi-modal public transport network and develops an algorithm to search for all feasible travel options given user-specified time constraints. 

Superviser: Rui Yao, Xuhang Liu

Students: Emilien Ulrich, Oreste Challandes

Although carpooling has been long celebrated as a promising solution to reducing vehicular traffic and emissions during peak hours, it is still limited within families and for long-distance trips. For instance, EPFL launched a carpooling platform to encourage its staff and students to carpool in their daily commute. However, limited trips are posted and very few users remain active. The inconvenient trip matching comes to be a major obstacle. In this project, we built an idealized carpooling system for daily commute trips that features flexible role assignment and centralized trip matching. 

Supervisor: Zhenyu Yang, Kenan Zhang

Students: Christina Schmid, Fannata Sédiko, Nicolas Wakim

“Braess Paradox” is a well-known phenomenon in traffic networks, where adding one road link to the network instead worsens traffic congestion. This discouraging consequence is rooted in the selfish routing behavior of private vehicles. For example, a newly built highway may attract a lot of people and finally intensify congestion on upstream and downstream roads. This project identified potential Braess links in the Sioux Falls network and explored their common properties. 

Supervisor: Hossein Farahani

Student: Julien Ars

This project analyzed the PubliBike network at EPFL and surrounding municipalities, including demand pattern and user satisfaction, using EPFL mobility survey and operational data. It also benchmarked current service against alternative solutions to identify key challenges and opportunities in promoting bikesharing. The project was led by the EPFL Mobility Office.

Supervisor: Luca Fontana, Luca Pellandini, Kenan Zhang

Student: Evangelia Gkola, Gregoire Ecuyer

This project studied the carpooling behaviors of EPFL staffs and students using the mobility survey and data of an EPFL carpooling platform. It also studied the implication of carpooling on EPFL’s parking and CO2 emission reduction strategies. The project was led by the EPFL Mobility Office. 

Supervisor: Luca Fontana, Alexia Couturier, Kenan Zhang

Student: Martin Simon, Marine Jacob

On-demand mobility covers a wide range of services including taxis, e-hailing (e.g., Uber), ride-pooling, and bike-sharing. Although different in detailed operations, these services share some fundamental characteristics regarding the movements of vehicles, their interactions with users, as well as the spatiotemporal distribution of demand and supply. This project investigated an open-source simulation of on-demand mobility and tested it potential for generalizing different service modes and market conditions.  

Supervisor: Kenan Zhang

Student: Mathis Magnin

The perturbed utility model (PUM) is a discrete choice model that represents an individual’s decision as a vector of choice probabilities and describes the utility of each alternative as the sum of systematic utility and a convex perturbation function of the choice probability vector. PUM has been shown to generalize a wide range of discrete choice models (e.g., MNL) meanwhile enjoying a higher computational efficiency in model inference. However, a challenge is the specification and estimation of the perturbation function. In this project, we explored the potential of learning the perturbation function using neural networks (NN).

Supervisor: Rui Yao

Student: Jingren Tang

Current navigation apps (e.g., Google Maps) usually recommend walking paths based on distance and elevation, while ignoring how much pollution the pedestrian may be exposed to along the walk. This project studied the pollution-aware path planning problem for pedestrians in the urban environment using the spatiotemporal environmental data collected by Sparrow.  

Supervisor: Rui Yao

Student: Anne-Valérie Preto