8-11 Mar. 2010 |
20-22 Jan. 2010 |
| In the field of satellites our main focus is on technology aimed at small low Earth orbit (LEO) satellites. Such satellites are of increased interest lately due to their lower development as well as deployment costs. As the name implies, LEO satellites fly close to the Earthfs surface; at altitudes of between 200km and 2000km. Below 200km, upper atmospheric gases cause too much drag, causing the satellite orbit to decay rapidly, and hence unstable orbits. LEO satellites travel at speeds of around 27,400 km/h (8 km/s), and complete one full revolution of the Earth in about 90 minutes. Due to their low altitudes and high speeds, LEO satellites used for Earth observation only cover small areas of the Earthfs surface, and for only short durations of time. Thus depending on the application, usually a number of these satellites are needed to provide the desired Earth coverage, leading to their deployment in groups so-called constellations. In order to gather real-time observation data from each of these satellites, it becomes necessary to first pass data between satellites in space, and then transmit it back to the Earth via the satellite flying over the ground station.
Our main goal here is to provide the technology that enables the realization of such a system in space. Due to their size, small LEO satellites have restrictions on computation resources as well as available power. In addition, due to their high orbital velocity, the nature of the network formed by these satellites in space is very dynamic. Also, the possibility of satellite failure exists and needs to be prevented from causing disruption to the application. All these various restrictions imply that a perfect solution to the problem must integrate various aspects of technology and it is what we are working on. We are happy to collaborate on technologies involved in solving the above problem, as well as general aspects of satellite and space technology. |
| The entry of microprocessors into the control loop of control systems has led to flexibility in such systems, and the realization of more advanced control loop algorithms. Though these devices are low cost and offer fast operating frequencies, their computation capability is limited by their basic architecture rather than their frequency. Processors are designed to execute instructions sequentially, which means that the primary way to enhance the control loop response of a processor-based control system is to increase the processor frequency. A type of processor that can perform better than the general purpose processor in a control system is the Digital Signal Processor (DSP). DSPs are specialized processors that work on the same principles as general purpose processors, but offer enhanced arithmetic computation capability. However irrespective of whether a general purpose processor or a DSP is used, there still exists a large array of control applications where both these processors fall far short of real-time requirements.
For such a class of control applications, where the processorfs sequential nature rather than its frequency are causing the real-time bottleneck, we suggest using an FPGA (Field Programmable Gate Array) instead of a fast processor. An FPGA might not be able to operate at the frequency of a fast processor, but it offers the capability of implementing parallelism in design, which can prove to be orders of magnitude faster, and also more reliable than a fast processor solution. Areas where we have successfully applied FPGAs to real-time control systems include multi-channel sensor data filtering, radar control and data acquisition, motor motion estimation and control, and real-time image filtering etc. To unleash FPGA on a real-time control problem, we can help get to the solution faster. |