
at Virginia Tech DARPA GloMo2 Final Report Introduction to Software Radio Technology Architecture Characterization and Design Computing Platform Organization and Design Software Radio Algorithm Modeling and Implementation Virtual Testbed for Simulating Performance Mapping the Receiver Algorithms onto Stallion Smart Antennas for Handsets Simulation and Modeling Technology Transfer Publications Last updated: January 10, 2001 |
Virginia Tech's GloMo2 Final ReportSimulation and ModelingKey accomplishments of the simulation and modeling efforts were (i) development of an OPNET model for the Virginia Tech CHARIOT radio, including the GloMo Radio API as its interface and integration with SAIC's SEAMLSS Lite, (ii) demonstration of the utility of the radio and of the model through simulation of a mobile ad hoc network (MANET), and (iii) demonstration of the proof of concept for hardware/software cosimulation of wireless networks with high fidelity link-level models.
Radio and Protocol ModelingThe simulation models consist of several components designed to work together in a direct sequence spread spectrum (DSSS) environment with a radio equipped with parallel interference cancellation (PIC) mode. Collectively, the simulation results show two- to three-fold improvements over traditional communication protocols. All models are either in an OPNET format or an OPNET-compatible format, making them easy to integrate with SEAMLSS Lite. The basic components developed are the CHARIOT radio model, a model of a new medium access control protocol called ADIMNB, and an associated MANET routing protocol called CAMEN. The CHARIOT model is an OPNET radio model with four-user multi-user receiver using a PIC algorithm. The model also reflects CHARIOT's ability to change spreading code, processing gain (PG), transmission power, and forward error correction (FEC) coding rate on a per-packet basis. The model captures CHARIOT's operating characteristics, including frequency use, preamble requirement, transmission range, and data rate. The adaptive DSSS medium access control (MAC) protocol for a non-broadcast multiple access medium (ADIMNB) is an opportunistic MAC protocol that takes advantage of unique capabilities offered by CHARIOT. It operates in an ad hoc network environment using spread spectrum multiple access (SSMA) as a primary mean of multiplexing, allowing CHARIOT to fully exercise its PIC capability. ADIMNB senses the environment and changes communication parameters, such as PG, power, and FEC rate, accordingly. It also encourages multiple simultaneous CDMA sessions in a half-duplex ad hoc network environment. The clustering algorithm for mobile ad hoc networks (CAMEN) is an extension of a table-driven ad hoc routing protocol. CAMEN improves network scalability by aggregating stations into clusters, which effectively reduces the number of routers in the system. This reduces the amount of routing information. Since ADIMNB prefers unicast messages, CAMEN also discourages the use of broadcast. CAMEN communicates a significant amount of environmental information and operating parameters with ADIMNB to further optimize the communication channel on a per-neighbor basis. ADIMNB optimizes the communication channel on a per-packet basis.
Interface Modeling and IntegrationThe GloMo Radio API is used extensively in both the ADIMNB model and the CHARIOT model. All communications across the radio interface are based on the GloMo radio API, including sending commands to the radio, reading and writing variables, and signaling. Using the GloMo radio API, ADIMNB can be used with other radio models or CHARIOT can be used with other MAC protocols. All models were integrated and successfully tested with the March 2000 version of SAIC's SEAMLSS Lite. The integration point is at the network level to provide routing capability to the radio. The integrated model uses a standard IP packet format on both sides of the interface. The integrated model, with constituent components indicated, is shown in the figure below. SEAMLSS Lite provides realistic traffic and movement patterns, and CAMEN, ADIMNB, and CHARIOT reliably deliver packets to proper destinations. Packet reception statistics and radio status are sent back to the SEAMLSS Lite modules for presentation and evaluation.
![]() Integration of model components with SEAMLSS Lite.
CosimulationThe OPNET radio and protocol models were successfully integrated with a link-level simulator implemented on a configurable computing hardware platform to demonstrate a novel cosimulation approach. The figure below illustrates the integration of the cosimulation method into network-level simulation. Only the link or links of interest are modeled using the hardware platform instead of all potential links in the network. Each transmitted packet is passed from SEAMLSS Lite to the hardware simulator interface and then to the hardware simulator. The packet reception result is then communicated back to the the OPNET model via the hardware simulator interface at the receiving end. Collisions, interference, and several types of noise are considered. Software components are fully integrated into OPNET radio transceiver pipeline, allowing easy integration with OPNET-based simulations such as SEAMLSS and SEAMLSS-Lite.
![]() System simulation using the cosimulation technique. The cosimulation approach allows complex protocols and networks to be simulated by commercial simulation software, offering realistic traffic and workload to link-level simulation studies. The method also allows straightforward but computationally-intensive channel models to be simulated using dedicated hardware, resulting in realistic link-level characteristics for network-level simulation studies. In addition, cosimulation has the potential to enable large-scale simulation by using multiple hardware platforms for a single simulation and to allow integration of the physical radio with the software simulator.
|