Si SDPS
The Si SDPS supports Air Traffi c Control operations for En-route (ACC), Approach (APP) and Tower (TWR) control. It was launched for operational use 1999 as a subsystem to the Si ATMSys and has been incrementally enhanced with new technique and functionality. The Si SDPS is in accordance with Eurocontrol criteria categorized as a very advanced ATM system.
Si SDPS has been designed for and proven to be
- Flexible and confi gurable for most sizes of ATC systems
- Easily expandable in size and functionality,
- Redundant in hardware and software
- Cost effective life-cycle
Input
Sensor data processing
10 sensors, extendable to e.g. 16
Formats: Asterix Categories 1, 2, 34, 48, 19 and 20
PSR, SSR and Combined radars, WAM and full Mode S
3D radars can be handled
Blanking areas for PSR and non-PSR plots can be defined
Plot Preprocessing
Handling of PSR, SSR and Combined Plots, WAM and full Mode S
Time stamping using plot, north or sector time
Conversion to internal coordinate system, WGS84, Transverse Mercator (exact Riemannian curved geometry; zweibein formulation)
Squint error correction
Triangulated height estimation in absence of Mode C
Bias compensation of plots, see below
Clutter map estimation for each radar
Single Radar Tracking
Parallel Kalman filters for uniform motion and manuevers. Stable design
Bias Calculation
Track correlator. Correlates single radar tracks belonging to the same target
Bias Estimation. Bias values are periodically calculated by system
Bias errors:
- Position Error
- Range offset and gain
- North alignment error
- Tilting antenna axis
- Time offset
Target dependent range error compensation, i.e. variation in test transponder delays
Multi Sensor Tracking
Track Initiation
Initiation by credibility; likelihood that track is real
Tracks stages, Possible, i.e. low credibility -> Tentative -> Firm, i.e. high credibility
Primary track initiation takes into account clutter density
Areas with special initiation criteria:
- Normal initiation
- Fast inititation
- Restricted inititation
- No initiation
Track-plot association
Maximum likelihood used to find associated track-plot pairs
Primary plot-track association takes into account clutter density, via Neyman-Pearson tests
Mixed-Gaussian pdf, used for describing the predicted track position, avoids track swaps
Handling to eliminated reflections and side-lobes
Fragmentation and formation handling
Track state estimation
Multiple model method used described by a discrete time Markov jump-linear system.
Tracking Algorithm: Generalized pseudo-Bayesian of second order (GPB2), which is identical to second orded Interactive Multiple Model (IMM2).
Kinematical models used:
- Uniform motion, Linear Kalman filter
- Coordinated Turn, Extended Kalman filter
- Uniform acceleration. Linear Kalman filter
- Turn with speed change. Extended Kalman filter
- Fast maneuvers. Linear Kalman filter
Ellipsoidal gating used to minimize effect of outliers
State vector: Position, velocity, linear acceleration, turn rate
Vertical Tracking
State update using recursive Bayesian estimation with z-measurement pdf realistically described by a uniform distribution, due to discreteness of Mode C
Exact state prediction of pdf by using Fokker-Plank equation
Output
Tracks are distributed periodically in batches in direction East-West or South-North
Direction and period of distribution are defined by parameters
Track output format: Asterix Category 62
System Capacity
500 System tracks (can be extended to 800)
400 Single radar tracks per radar
Technical Terminal
Parameter handling
Online update of parameters
System Monitoring and control
Track and plot load, bias values, cpu load
Error logging
Technical Features
Modern open system architecture
Commercial hardware
- UNIX workstations
- Fault-tolerant servers
- Redundant LAN
Standard protocols are used and the application software is in C, C++ and ADA
Unix operating system
- Platform dependant, AIX, DEC Unix, Lynx OS, LINUX
- X Windows / MOTIF
Working position equipment
- Working positions with dual or single displays, selectable in size and resolution
- High-intensity TWR monitor
- Standard keyboard and mouse
General Design aspects
- Client/server concept
- ADA, C and C++
- Distributed processing
- Fault-tolerant and/or redundant hardware
- Fault-tolerant software
- Non-dedicated workstations
System Redundancy Concept
The system can be delivered with various redundancy levels:
- Hardware redundant on server level and on network level, through the use of fault tolerant computers and dual networks
- Redundancy in data communication between sensors and SDPS through serial lines and network connections
- System redundancy on hardware and functional level, through processing blocks in Master/Standby configurations
- Final level of redundancy might be a separate Radar ByPass System with own input channels and connections to users through separate paths
Landsvägen 39, 172 63 Sundbyberg, Sweden