Contact Person
Ilja Radusch
Dr.-Ing. Ilja Radusch
Director
Business Unit ASCT
+49 30 3463-7474
Dr.-Ing. Robert Protzmann
Business Unit ASCT
+49 30 3463-7289
Eclipse MOSAIC
Fraunhofer FOKUS

Smarter Mobility with Simulation

Open Source simulation environment Eclipse MOSAIC

Smart Mobility applications and Cooperative Transportation Systems help to enhance safety and traffic efficiency. However, the simulation of corresponding scenarios is a challenge because different simulation worlds come together here: e.g. vehicular traffic, wireless network communication, and application modeling. To solve this problem, MOSAIC couples different simulators and enables the simulation of the various aspects of future Intelligent Transportation Systems. Consequently, MOSAIC simulations allow the analysis of novel mobility solutions before real field tests start.

MOSAIC is a multi-scale and multi-domain co-simulation framework for the assessment of new solutions for connected and automated mobility. The Open Source version Eclipse MOSAIC offers everything you need to assess smart mobility scenarios, while the enhanced version MOSAIC Extended provides further tools and models, such as the vehicle dynamics simulator PHABMACS.

Access to MOSAIC

The MOSAIC simulation framework comes in two flavors: the Open Source version and an extended commercial version. Eclipse MOSAIC is Open Source under the umbrella of the Eclipse Foundation and the complete source code can be downloaded from Eclipse Github. For a quick start we provide a pre-configured bundle. This bundle includes the basic simulators and tools required for the assessment of novel connected mobility applications with vehicle-2-x communication.

MOSAIC Extended is the commercial version of MOSAIC and comes with additional tools and simulators. This bundle can be customized so that it fits perfectly to your needs and use-case. A quote of the MOSAIC Extended bundle depends on the additional compontents you need, which are, amongst others, simulators for electric mobility, the MOSAIC Statistics Generator, a 3D visualization, and the PHABMACS vehicle simulator.

Co-Simulation

In contrast to existing fixed simulator couplings, the MOSAIC simulation infrastructure allows the easy integration and exchange of simulators. Thus, the high flexibility of MOSAIC enables the coupling of the most appropriate simulators for a realistic presentation of vehicular traffic, emissions, wireless communication (cellular and ad-hoc), user behavior, and mobility applications. All simulator management tasks, such as time and data synchronization, interaction and lifecycle management, are handled completely by the MOSAIC runtime infrastructure (RTI).

Multi-Domain

Simulators usually cover specific domains, such as traffic or communication. With MOSAIC one can combine the most suitable simulator of each domain to build a holististic system simulation. The following simulators are already coupled with MOSAIC: Eclipse SUMO, our PHABMACS simulator (Traffic), OMNeT++, ns-3, MOSAIC Simple Network Simulator (SNS), MOSAIC Cell (Communication), MOSAIC Application (Application and behavior), and more.

Multi-Scale

Depending on the specific requirements of a simulation scenario, the most relevant simulators can be used. Even within a specific domain different simulators model different levels of domain. For example, traffic simulations can handle thousands of vehicles at once, but only on a low-detailed modelling level which might not be sufficient for specific use cases. For this purpose, a vehicle simulator modelling vehicle dynamics on a high-detailed level might be used, such as PHABMACS. With MOSAIC the exchange of these simulators on different modelling levels is uncomplicated, allowing simulations of advanced driver assistence systems (ADAS) of single cars, up to analyzing the traffic efficiency of Smart Mobility applications on a city-wide level.

Modelling and Configuration

The MOSAIC Application simulator is a simple way to test and evaluate Smart Mobility applications with MOSAIC. Applications for different entities, such as vehicles, servers, and road side units, can be integrated easily thanks to a well-established Java-API. Custom applications can be mapped onto all entities using a sophisticated mapping configuration, allowing easy adjustments of penetratration rates and traffic volumes. Following the Seperation-of-Concerns principle, each simulator is configured separately allowing a flexible parametrization of simulation scenarios.


Related Links:
asct, phabmacs
Fraunhofer FOKUS

Prototyping Cooperative ADAS with PHABMACS

Physically realistic simulation allows for quick prototyping of cooperative Advanced Driver Assistance Systems (ADAS) for vehicles and VRUs. PHABMACS can simulate complete virtual vehicle environments including vehicle dynamics, sensors, HD lane accurate road infrastructure and communication elements.

As an cost-efficient alternative to commercial vehicle simulators, PHABMACS is designed to be a lightweight, developer-friendly, easy to modify tool to prototype cooperative driver assistance systems. Developers can use the simulated vehicles in PHABMACS first, before applying there code in real test vehicles.

Respective interfaces for sensors and actuators of the simulated vehicle models are provided to the prototype system and can later easily exchange with real sensors and actuators. Additional to a wide range of radar and lidar sensors, communication-as-a-sensor between vehicles and infrastructure is provided to enable cooperative system functions.

Application code and additional sensor models can either be created in JAVA™ and deployed into the PHABMACS framework or interfaced from MATLAB/Simulink®. In combination with our CAN bus hardware interface, PHABMACS can also be hooked up in real-time with any vehicle controller setup and replace real-world vehicles for testing.

Scope

Fitting the scope of prototyping cooperative Advanced Driver Assistance Systems, PHABMACS addresses two conflicting requirements:

  • Driver assistance systems need to be aware of the driving dynamics
  • Cooperativeness requires as many vehicles as possible being part of the simulation

For the former purpose, the simulation needs to map physics as realistic as possible, which results in high computational effort per vehicle. For the latter, we optimized PHABMACS the computational effort by mapping physics realistically, when vehicle parameters are below the limits of driving dynamics, which is sufficient for its scope. Regarding the maximum number of vehicles in the simulation and its level of detail, PHABMACS closes the gap between traffic simulators and high detail vehicle simulations like FEM, K&C, or MKS.

Vehicle Model

The applications to be prototyped can access the simulated vehicles in PHABMACS by controlling their actuators (throttle, brake, steering) and read their sensors. The input on throttle and brake are transformed by the power train (and brake) to move the vehicle model. The powertrain include separate modelling of the engine, the torque converter (and clutch alternatively), the transmission and the wheels. The interaction of the vehicle and other objects of its environment is based on rigid body dynamics.

Environment Model

The environment in PHABMACS allows for creaion of arbitrary scenarios. Conveniently, such scenarios can be automatically generated from OpenDRIVE® and Open Street Map (OSM) map material using the street grid, buildings and natural elements like trees, grass, and water. Additional infrastructural elements like construction sites can be added. The standard visualization theme has a puristic, yet appealing rendering style. Customized themes can be implemented easily.

Part of MOSAIC simulation framework

The integration with MOSAIC Extended enables coupling PHABMACS with other simulation tools of different domains, such as network simulators for V2X/5G communication, specific sensor simulation models, or application modelling tools. For example, simulating V2X communication using the single hop communication models of PHABMACS, is restricted to map simple metrics like delay and packet loss. If a complex communication scenario is needed, e.g. OMNeT++ or ns-3 can be coupled. Combined with traffic simulation tools like SUMO, the PHABMACS simulation can be embedded in large scenarios of communicating vehicles.