Automatic evaluation of air mission routes with respect to combat survival
Graphical abstract
Highlights
- •
A survivability model is suggested that describes the risks for flying a route.
- •
The model captures the dependency between getting tracked and getting hit.
- •
An expected cost measure for route evaluation is suggested.
- •
Simulations show the influences of uncertainties regarding the enemy’s locations.
Introduction
A fighter pilot flying a mission inside hostile territory is exposed to the enemy’s ground-based air defense systems. The best way of surviving the mission is to fly outside the enemy’s weapon ranges. However, this is not always possible. First of all, the enemy locates weapons to protect their valuable assets and hinder the fighter aircraft to reach them. The pilot therefore needs to balance the objectives of flying the aircraft in a safe way, accomplishing the mission and combat survival, i.e., flying the route unharmed [1]. Secondly, information regarding the enemy’s sensors and weapons is often uncertain, since the enemy camouflages its systems and relocates them frequently. Even though the pilot plans the route with the best available intelligence information, the situation might have changed when the mission is flown. The pilot therefore needs to analyze the situation when new information is received and re-plan or abort the mission if the situation is too dangerous [2]. The purpose of this work is to suggest a method for automatic evaluation of mission routes from a combat survival perspective. This can be used for planning the mission by comparing several possible routes and selecting the least dangerous one. During flight, it could aid the pilot to re-evaluate the route when updated information regarding the enemy systems is received.
The enemy’s air defense systems consist of both sensors and weapons system [3], which are potential threats to the aircraft and should be avoided if possible. The danger posed by the enemy’s defense systems depends on the enemy’s capabilities and intentions to harm the aircraft. Furthermore, the enemy’s opportunities to detect and hit the aircraft depend on the aircraft’s position as well as the previous part of the route. For instance, if the aircraft has flown inside the range of the enemy’s sensors for a long period of time when it enters a weapon area, the risk that the enemy fires a weapon against it is high. On the other hand, if the aircraft can enter the weapon area undetected, the survivability, i.e., the probability of flying the mission unharmed, is higher. The evaluation of the route is further complicated by the fact that information regarding the enemy’s capabilities, positions and intentions are typically held secret and therefore needs to be estimated based on intelligence information and sensor data. This information is therefore uncertain and the evaluation of the routes needs to take this uncertainty into account. Hence, for assessing the survivability for a mission, information regarding the enemy’s capabilities and positions should be fused with information regarding the history as well as the future of the mission route. This corresponds to Level 3, Threat Assessment, in the JDL model, since the aim is to predict the enemy’s intentions, capabilities and future opportunities to harm the aircraft [4], [5].
Models for describing the risk or danger associated with flying a route have been suggested in the literature regarding decision support for fighter pilots and operators of unmanned aerial vehicles (UAVs) as well as in the literature for route planning in hostile environment. The most straightforward approach is to consider the enemy’s defense systems as no-fly zones that should be avoided [6], [7] or to minimize the time inside these zones [8]. However, when the mission requires that the aircraft flies near the enemy’s systems, there is a need for more complex models for determining where to fly. More advanced approaches focus on how the aircraft can use the terrain to hide and describe the time the aircraft is visible for enemy sensors [2], [9]. Inspired by the radar equation, the danger associated with a part of the route passing near a radar station has been described as proportional to , where d is the distance to the radar, see e.g., [10], [11], [12], [13].
The approaches above all used deterministic descriptions of the relation between the aircraft and the enemy’s systems. However, the information regarding the enemy’s intentions and capabilities are typically uncertain and several probabilistic modeling approaches have therefore been suggested. Berger et al. [14] calculated the probability of flying a route unharmed and Pfeiffer et al. [15] calculated the probability that a UAV could fly a route undetected or detected at most k times. Detection could either be interpreted as a fatal attack, in which case no detection is desirable, or as detection by enemy sensors, in which case can be acceptable. Kim and Hespanha [16] presented a similar model, which described the probability that a group of UAVs could fly a route undetected given their use of radar jamming. Kabamba et al. [17] argued that the UAV must be tracked by radar for some time before a weapon can be fired and also during the time it takes for the missile to reach the UAV. They therefore modeled the probability that the enemy radar is able to continuously track a UAV during a sufficiently long period of time. Other references have focused on the enemy’s weapons and the risk of getting hit. Dogan [18] utilized a probability density function for describing the probability that the aircraft would get hit at a position. Randleff [19] suggested a similar approach where each position was associated with a threat lethality affected by the pilot’s use of countermeasures. More detailed models have also been proposed. Ögren and Winstrand [20] used simulations with surface-to-air missile models. Theunissen et al. [2] suggested a weapon state diagram for calculating the probability that a weapon hits the aircraft by identifying the underlying events and their probabilities.
The enemy possesses both sensors and weapons and they are all potential threats to the aircraft and affect its chances of flying the mission unharmed. It is therefore suitable with models that consider both the probability of detection and the probability of getting hit. Xinzeng et al. [21]and Besada-Portas et al. [22] suggested probabilistic models for radar and weapon systems, but handled these threats as independent. On the other hand, several references have described the risk at a position as the combination of the probability that the aircraft is tracked and the conditional probability that the aircraft gets hit if it is tracked, see [23], [24], [25]. It is assumed that these probabilities at one position are independent of where the aircraft has previously flown. This approach can be suitable if each sensor is co-located with a weapon and the different sensors are assumed to work independently. However, in an integrated air defense system, several sensors are searching the airspace and share information with each other and with the weapon systems, which not necessarily have the same range as the sensors. Furthermore, as commented above, the aircraft must be tracked by radar for some time before a weapon can be fired. The probability that the aircraft will get hit inside the range of the weapon therefore depends on how long time it has previously flown within the range of the sensors. The references identified in the literature review are not able to capture this dependency, but the model suggested in this paper has been developed to allow for this.
The literature reveals different approaches for handling uncertainty regarding the positions of the enemy’s systems. A straight forward approach is to ignore the uncertainty and assume that the position are known, see e.g., [17], [22]. Another approach is to construct a probability map, where each position on the map is associated with the probability that an enemy system occurs at that position. The probability map can either be assumed to be known, see e.g., [14], [19] or constructed based on surveillance information see e.g., [24], [25]. Yet another approach was used in [23], where the threats’ positions were regarded as random variables with known distributions. These distributions were used for calculating the corresponding distribution of the survivability of the route. This latter approach will be utilized also in this paper, since it has the nice property of separating the uncertainty that is associated with the capability and intents of the enemy and the uncertainty that stems from the locations.
Automatic evaluation of routes from a combat survival perspective requires a model of how the enemy’s systems affect the aircraft opportunities to fly the route unharmed. Discussions with domain experts reported in [26] resulted in the following assumptions regarding the enemy’s systems.
- •
The enemy has access to both sensors and weapons, which are not necessarily co-located.
- •
The enemy’s system can share information.
- •
The aircraft must be tracked during some time before a weapon can be fired.
- •
The aircraft is much faster than the enemy’s systems and these can therefore be modeled as stationary.
Contrary to the models identified in the literature review, the survivability model presented here incorporates all these assumptions. It explicitly describes both the probability of getting tracked by the enemy’s sensors and the probability of getting hit by the enemy’s weapon as well as the relation between these events. An initial implementation of the model was presented in [27], [28]. This paper extends the model and the analysis of its usage by:
- •
Extending the model to handle the case when the sensor and weapon system are located such that their ranges overlap, see Section 2.
- •
Introducing and analyzing a method for evaluating routes based on the survivability model, see Section 3.1.
- •
Analyzing and incorporating the influence of position uncertainty regarding the enemy’s systems in the evaluation of the route, see Section 3.2.
Section snippets
Survivability model
The survivability model is a continuous-time Markov model where the states describe the relationship between the aircraft and the enemy. Markov models are used for describing stochastic processes in a number of domain, e.g., target tracking and classification [29], reliability theory [30], life and sickness insurance [31] and medical decision making [32], [33]. The description here is based on [34], [35].
Evaluation of routes
The aim of the survivability model is to evaluate a route or to compare different routes in order to plan where to fly. However, how such a comparison should be performed is not trivial. Plotting the state probability vectors over time gives a rich description of the routes, as was shown in Fig. 3. This can be used for identifying critical parts and assessing the dependency between the probability of getting hit and the probability that the aircraft is detected and tracked. The pilot can use
Conclusions and future work
Aircraft flying in hostile environments are exposed to the risk of getting hit by enemy fire from ground-based air defense systems. It is not always possible to both accomplish the mission and stay outside the range of the enemy’s weapon systems, especially if the positions of the enemy’s systems are not perfectly known. In such cases, the pilots could be aided by a system that automatically evaluates different routes from a combat survival perspective in order to determine which route to fly.
Acknowledgements
This research has been supported by the Swedish Governmental Agency for Innovation Systems (Vinnova) through the National Aviation Engineering Research Program (NFFP5-2009-01315), Saab AB and the University of Skövde. We would like to thank the three anonymous reviewers for their valuable comments.
References (37)
- et al.
Cognitive situation and threat assessments of ground battlespaces
Inform. Fusion
(2003) - et al.
A new mixed-integer linear programming model for rescue path planning in uncertain adversarial environment
Comput. Operat. Res.
(2012) - et al.
On target classification using kinematic data
Inform. Fusion
(2004) Cognitive automation for tactical mission management: concept and prototype evaluation in flight simulator trials
Cognit., Technol. Work
(2002)- E. Theunissen, F. Bolderheij, G. Koeners, Integration of threat information into the route (re-)planning task, in:...
- F. Johansson, Evaluating the Performance of TEWA Systems, Ph.D. thesis, Örebro University, School of Science and...
- et al.
An introduction to multisensor data fusion
Proc. IEEE
(1997) - A. Goossens, G. Koeners, E. Theunissen, Development and evaluation of level 3 situation awareness support functions for...
- B. Rosenberg, M. Richards, J. Langton, S. Tenenbaum, D. Stouch, Applications of multi-objective evolutionary algorithms...
- et al.
Comparison of parallel genetic algorithm and particle swarm optimization for real-time UAV path planning
IEEE Trans. Indust. Inform.
(2013)
Coordinated target assignment and intercept for unmanned air vehicles
IEEE Trans. Robot. Auto.
Evolutionary route planner for unmanned air vehicles
IEEE Trans. Robot.
Probabilistic modeling for UAV path planning in the presence of threat zones
Optimal path planning for unmanned combat aerial vehicles to defeat radar tracking
J. Guid., Control, Dynam.
Cited by (29)
Feedback RRT* algorithm for UAV path planning in a hostile environment
2022, Computers and Industrial EngineeringCitation Excerpt :The cooperation of radar and weapon systems to attack a UAV is a dynamic process, and the final destruction of a UAV is usually caused by multiple events such as radar detection, missile launch, and missile hit (Pınar and Koyuncu, 2018). Based on assumptions about enemy systems (Erlandsson and Niklasson, 2014). the characteristics of the defense process considered in this study include the following.
Multi-UAVs trajectory and mission cooperative planning based on the Markov model
2019, Physical CommunicationCitation Excerpt :Moreover, while guaranteeing the highest total effect value, the survival probability of the aircraft cannot be guaranteed, but the survival probability of UAV will affect the total effect value of mission execution in turn. As for these problems, the Markov state model mentioned in the literature [23] and the state transition and cost calculation in the literature [24] are introduced. To grasp the specific survival probability of the UAVS during flight, on which the search is modified to match the particle swarm algorithm [25].
Image quality tendency modeling by fusing multiple visual cues
2019, Journal of Visual Communication and Image RepresentationDecision support system for optimum decision making process in threat evaluation and weapon assignment: Current status, challenges and future directions
2017, Annual Reviews in ControlCitation Excerpt :In real-time environment, TE process is considered challenging under normal conditions and possibly worse under extreme situations due to limited time, stress and multi-tasking demands. Ultimately, it is the responsibility of human operators (Berthelemy et al., 2002; Daniel, 1979) to perform data analysis and propose recommendations on basis of data provided for TE thus entailing high risk and complexity (Deng et al., 2010; Erlandsson & Niklasson, 2014; Naseem et al., 2014; Riveiro et al., 2014). It may propagate uncertainty and risk making it imperative to reduce the cognitive load on operators while enhancing threat assessment and decision making (Azimirad & Haddadnia, 2015; Riveiro et al., 2014).
Dynamic Adversarial Risk Estimation Based on Labeled Multi-Bernoulli Tracker
2024, Journal of RadarsSensitivity of Single-Pulse Radar Detection to Aircraft Pose Uncertainties
2023, IEEE Transactions on Aerospace and Electronic Systems