Contact Us

Use the form on the right to contact us.

You can edit the text in this area, and change where the contact form on the right submits to, by entering edit mode using the modes on the bottom right. 

1 Create Way
Singapore, 138602

Research project within TUM CREATE. Focussing on modelling and optimisation of architecture and infrastructure, urban systems simulation like traffic and power are the main research interests. Apart from that, a cognitive systems group deal with human computer interaction. 


A Simulation-based Heuristic for City-scale Electric Vehicle Charging Station Placement

Jiajian Xiao

Electric Vehicles (EVs) play an important role towards a more sustainable transportation system. Sufficient charging infrastructure is, however, needed in order to accommodate their power demand and increase EV adoption. In this study, a simulation-based approach for charging station (CS) placement using an agent-based traffic simulation is proposed. The heuristic's objective is to achieve sufficient network coverage to keep charging related inconvenience within an acceptable range while minimising the overall number of CSs. For this purpose, the algorithm identifies locations at which the charging procedure seamlessly integrates into the drivers' itineraries, thus minimising detours and waiting times. At the same time, the algorithm attempts to maximise the utilisation of each CS throughout the day in order to minimise the number of CSs. The methodology is demonstrated at the example of Singapore. The investigation shows that the charging demand of 20,000 EVs can be covered with approximately 2,500 CSs by accepting average detours no greater than 410 metres and average waiting time below 10 minutes. This number can be further reduced by relaxing the inconvenience criterion. 

The main components of the proposed methodology are a charging behaviour model and a CS placement algorithm. The first one determines under what conditions a driver decides to recharge the vehicle’s battery, the latter aims for placing CSs in a way which best suits the drivers’ charging needs.

For charging behaviour, two types of charging behaviours are defined depending on the battery usage:

  • Mandatory Charging: charge when battery is not sufficient

At the end of each trip, an estimation is made to check if the remaining energy is enough to support the next trip + some safety margin.  And mandatory charging is triggered when the remaining energy is less than the estimation

We introduce a queuing behaviour. When a mandatory charge is required and the CS is currently occupied, a vehicle will queue up for charging. We assume an unlimited queue for each CS.

  • Convenience Charging: charge during parking

Convenience charging is modelled by considering a trade-off between remaining the energy and the distance to the nearest CS. 

CS placement algorithm removes unused CS and merges nearby CSs with less usage.

The combined usage of two CSs is computed by the following formula. If the combined usage is below a certain threshold, these two CS will be combined.


And the new CS will be created at the place which uses the idea of centre of mass

We run the experiment with CityMoS. The best result we achieved is:

With 20,000 EVs, we achieve a distribution of CSs as illustrated by the figure. 

  • Approximately 2,500 CSs
  • Average detour 410 meters
  • Average queuing time of 10 minutes
Screen Shot 2017-12-21 at 15.51.20.png

CityMoS 1.6 - Nasi Lemak

Daniel Zehe

CityMoS 1.6 is the third large feature release of CityMoS. In the tradition of naming CityMoS releases after items from the Singaporean cuisine, this release is called Nasi Lemak.

In addition to bug fixes and overall improvements to the stability and usability, this release introduces the ability to add custom dwelling time models for the busses, as well as an improved traffic assignment model. This optional feature improves the route choice model of vehicles in order to resemble human decision making then navigating through a traffic system.

In addition to the new features in CityMoS Nasi-Lemak, improvements to existing functionality were added. This includes the introduction of shape points for individual lanes. This allows modeling the curvature of a lane independently of the geometry of the associated Link. One effect of this alteration in road geometry modeling is a significant reduction in overall network size and complexity. Before, additional road links had to be introduced to resemble the curvature of a road. Appart for a reduction in network file size, the routing performance due to reduced graph complexity can be improved. A modularization of the lane changing allows for custom lane change behavior models to be implemented in the future. 

The CityMoS 3D front-end introduced with CityMoS 1.5 has improved road texture renderings as well as traffic signal models at intersections. Custom themes for styling the visual output can be added either manually or by using the CityMoS Configuration editor.

CityMoS with Car Models.png


The tooling for creating experiments with the CityMoS Configuration Editor and creating and modifying networks has been updated to versions 1.4 and 1.5 respectively. The CityMoS Network Editor supports the editing of the newly introduced lane shape points as well as better keyboard hotkey support for quickly changing between different tools.
CityMoS Configuration Editor supports multiple bus lines as well as general usability improvements to the workflow to work with the changed input file structures for CityMoS 1.6. It also offers the user the possibility to create configurations that work with CityMoS 3D.

PhD Thesis: Load Balancing and Synchronisation for Parallel Agent-based Road Traffic Simulation

yadong xu

In June 2017, Xu Yadong obtained his Ph.D in the School of Computer Science and Engineering at Nanyang Technological University. His PhD topic focuses on novel algorithms to accelerate the execution of microscopic traffic simulation using high performance computers. This is particularly useful in the studies where real-time decision making is critical, or a large number of simulation runs are required.

Read More

How will the charging of electric vehicles affect Singapore’s power grid?

Vaisagh Viswanathan

The large-scale introduction of plug-in electric vehicles (PEV) may pose challenges to power system operators by causing grid congestion or voltage fluctuations. In a previous post we presented a simulation framework consisting of a traffic simulation coupled to a power system simulation allowing investigating the impact of electric vehicle charging on the power grid. In this post we present a first application of this approach to the case of Singapore.

Read More

Modelling of qualitative linguistic spatial terms for Human-Robot Interaction

Vaisagh Viswanathan

In our group we address the symbol anchoring problem by (1) exploring mixed symbolic and geometric representations that allow a robot to integrate information coming from different sources (verbal communication, sensors, background knowledge, etc.), (2) development of geometrical models of spatial terms as qualitative spatial relations or route instructions. This blogpost focuses on the later.

Read More

First glimpse of the effect of uncertainty on charging behaviour

Vaisagh Viswanathan

As an application to the previous article, we examine how uncertainty influences charging behaviour. Due to limited range and long charging time of electric vehicles (EVs),  the range anxiety phenomenon occurs in situation where the user might not reach the destination when low on battery (M. Nilsson 2011). While internal and personal factors are important factor regarding range anxiety (N. Rauh et al. 2014), we focus on external factors. Understanding charging behaviour will enable more optimal deployment of charging stations. In this article, the effect of uncertainty on charging behaviour is addressed using  a serious games approach or a game-based experiment.

Read More