Angle Up

Operator Performance

Human-machine interaction in future AvSec screening

Continuous development of Artificial Intelligence (AI) is part of the road towards future baggage screening in aviation security. Explosive Detection Systems for Cabin Baggage (EDSCB) aid operation in the detection of bare explosives and continue to improve performance in terms of detection rates. AI techniques, such as machine learning and deep learning, enable the development of smart and adaptable algorithms for automatic detection of threat objects - referred to as Automatic Prohibited Item Detection System (APIDS).  These algorithms enable support in security operations by detecting threat categories by shape, such as sharp items and firearms.

In this blog we take a look at the road towards future baggage screening in airport security, specifically at the human-machine interaction between operator and algorithm.


Continuous improvement of EDSCB algorithms

The continuous improvement of EDSCB already evolved to a single digit false alarm rate and an improved detection rate. Improvements of EDSCB detection can be related to the detection of minor threat quantities and threat concealments. One key element missing in current algorithms is the detection of complete Improvised Explosive Devices (IED).   

When EDSCB algorithms are able to detect IED’s, this will highly impact the security operation since operators hypothetically won’t have to search for IEDs anymore. As we described earlier in our blog about Operator performance on image analysis, the search for IEDs is the most complex detection category for operators. Because the threat category list is reduced by IEDs in operating procedures, the operator learning curve could increase. In addition, less operator training and coaching could be necessary.

Reducing the threat list with IEDs can result in an enhanced focus on other threat categories and this will most likely impact the operator satisfaction and confidence. The operational impact could be an increased clear baggage flow and fewer high threat occurrences; this expects to minimize the so-called Cry-wolf effect: ignoring warnings due to the past experience of false alarms.

Figure 1 – Operators must search for prohibited items in airport security


Automated Prohibited Item Detection and Image On Alarm

Automated Prohibited Item Detection Systems – APIDS – are the next step towards a more secure and automated security check. These advanced algorithms are able to detect and identify threat items such as firearms, sharps and blunt weapons, based on shape.

Both EDSCB and APIDS detect threats. Yet – their detection methodologies are significantly different. EDSCB detects bare explosives primarily by material composition, whereas APIDS detects threats mostly based on shape. On Screen Resolution (OSR) of EDSCB alarms is not allowed in many countries because the human eye is incapable of material distinguishment by observation. On the contrary, operators are able to perform shape detection, which is applicable on all threat categories, except explosives (see figure 2). Therefore, a so-called Image On Alarm (IOA) is an applicable CONOP scenario for APIDS: operators should only analyse images that contain a threat detected by APIDS.

Figure 2 – Material detection on explosives versus shape detection on firearms, sharps, blunts and miscellaneous

Successful implementation of APIDS requires consideration of different CONOPs scenarios. A CONOPs scenario with IOA allows for alarm resolution by the operator and could impact the operation on several aspects. Fewer images will be presented to the primary screener which results in an additional increased clear flow of baggage. Next to this, an alternative staffing model could be considered as fewer images requires analysis by an operator: remote screening with fewer operators or combining CBS and HBS operator activities.

When images are presented with alarm annotation operators can adjust their focus to a specific region of interest instead of analysing the complete image; operator focus will increase, and analysis time will decrease. Yet, it is important to consider additional operator GUI interaction due to image clutter of alarm annotations. Changes to the Operator GUI could help to increase intuitive GUI usability.

Fewer false alarms will be sent to recheck for secondary screening. Passenger experience could increase because this results in less queue time and less intrusive procedures. This also means an increased number of correct alarms that require secondary screening. Secondary screening on correct alarms requires more processing time than secondary screening on false alarms.


The challenges and opportunities of Auto-clear

Combining EDSCB with APIDS is the next step towards a more automated security checkpoint in the future. When the detection threshold enables the detection of all threat categories this will allow for implementation of auto-clear software; images can be cleared automatically if no threat is detected by EDSCB and APIDS. The implementation of auto-clear will impact the operational environment on various aspects.

Implementation of auto-clear software will result in fewer rejects of images. Therefore, operators will shift their focus towards alarm resolution procedures. The initial and recurrent training program for image analysis should then be adjusted accordingly.

The staffing model will be impacted by auto-clear implementation as well. Performing less alarm resolution also enables the opportunity to combine operator tasks. For example: combining passenger screening with alarm resolution for cabin baggage, or alarm resolution with first-line equipment support.

Rotating shifts within the security team on site will change with an auto-clear CONOP. There will be less start-stop behaviour of the lane and a passengers will experience a more continuous flow.

Auto-clear software will help in the improvement process of the security culture. Operator confidence and operator satisfaction levels might increase, different career opportunities arise, passengers will experience a less intrusive security procedures atmosphere and less need for ethnic profiling might exist.

Auto-clear will also impact the operational environment physically. Security lanes could require extended reclaim space for passengers when auto-clear software is implemented.  Divest positions could be extended as well due to a more continuous flow. Also, the position for the primary screener is no longer needed in the security lane. Therefore, this space becomes available. This could allow for narrow security lane setups, improvement in equipment sustainability and a more efficient equipment utilization. However, when changing the checkpoint footprint, the potential displacement effect of bottlenecks in the passenger journey – such as border control processes – must be considered as well. Relocating the bottleneck is never the objective in the passenger journey and must be considered in alignment to auto-clear implementation.


In this blog post we shared our thoughts on the future of AvSec Human-Machine interaction and how this can impact several factors in airport security. Dialogues are necessary among all European security stakeholders to define AI operational requirements, perform the research and guide the innovation process to strive for state-of-the-art AvSec solutions compliant with Europe’s regulations.

We guide security equipment deployment by keeping integration as our key focus. This covers technology, process and at heart: people. Every change results in a new situation.


Explore what Point FWD has to offer in guiding security stakeholders toward successful equipment implementation in airport security checkpoints.

Visit our EDSCB page for more info or send an information request via the button below.

The impact of operator satisfaction on performance levels.

There are various factors that affect the performance of security operators during the image analysis task. We discussed three of those factors in our previous blogs: Human perception, Checkpoint Environment and Legislation & CONOP. Yet, there is another important factor that cannot remain undiscussed: Operator Satisfaction.

A decent amount of research has concluded on the fact that a higher job satisfaction causes a higher motivation to do work, and therefore a higher job performance. We reached out to three AvSec field experts who shared their thoughts, being a valuable contribution in our discovery of the link between operator performance and operator satisfaction.

Brian Cilinder-Hansen Security Process Manager

Brian Cilinder-Hansen
Security Process Manager

Sofie Eyckerman  Aviation Quality & Training Manager

Sofie Eyckerman
Aviation Quality & Training Manager

Maykel Bloom  Security Coordinator

Maykel Bloom
Security Coordinator

The most important aspects that influence operator satisfaction, according to the field experts, can be scaled under the following: work environment & equipment, training & coaching, and operator involvement & personal motivation.


Work environment and equipment

As mentioned in one of the previous blogs, there are various screening configurations which enable different screening environments for operators. How a certain environment impacts an operator is dependent on the operators themselves because every operator is different. Some operators might find an on-site environment distracting and can be more satisfied screening in a quieter remote location, while others work well with the checkpoint commotion as background noise. Regardless of what configuration is applied, the screening environment should not distract the operators from their responsibilities on image analysis. The field experts elaborated on this.

MAA - PMT-03.png
With a comfortable
working environment and supportive equipment, the job should become easier which leads to a higher satisfaction and performance.
— Maykel Bloom PMT Security Coordinator Maastricht Aachen Airport

Regarding operator satisfaction, Sofie Eyckerman (G4S) mentions, subtleties like a blinking screen or too bright lighting are all to be considered impactful. She adds: “Technical aspects are important as well. If, for example, the functionality buttons (of the image analysis user interface) don’t work properly or the image representation is not right, the screener will not be able to perform their analysis to the best of their capabilities. Which can be very frustrating.” This could cause a bottleneck that is not only frustrating for the operators, but for the airport as well given that their security performance depends on this.

Brian Cilinder-Hansen (CPH) mentions that Copenhagen Airport is currently working towards Centralized Image Processing (CIP) – also known as remote screening – which will allow for a pool of expert screeners. He explains: “It’s important to consider operator satisfaction in this process, because we should create an environment that works best for them, while at the same time ensuring that we meet all security standards. We must find the right screening environment for both CBS and HBS screeners here in CPH.”

Maastricht Aachen Airport recently upgraded their checkpoint from a single view system to dual view with EDS and re-check system. Maykel Bloom (PMT) gives some insight on the effect of providing operators with this new and upgraded equipment. They saw a significant increase in the satisfaction and performance of their operators. He explains: “With a comfortable working environment and supportive equipment, the job should become easier which leads to a higher satisfaction and performance.” Training and coaching methods allow operators to gain the knowledge and skills needed to use this equipment.


Training and coaching

G4S-02.png
Sufficient training is a foundational condition for finding the confidence needed for this job.
— Sofie Eyckerman Quality & Training Manager G4S Belgium

Before operators are allowed to perform image analysis in practice, they must be trained. As Sofie Eyckerman (G4S) says: “Sufficient training is a foundational condition for finding the confidence needed for this job”. But the definition of ‘sufficient’ can vary between operators. Maykel Bloom (PMT) gives the following example: “Recognizing items in an X-ray image can be very intuitive and logical for one person, while another person might need to physically see an item before being able to recognize it in an image. When operators discover what works for them specifically, this adds to their satisfaction.” Even though it would take some time, investing in a personalized training method could be beneficial for both the operators and the security company / airport. It could result in a more effective and time efficient learning curve.

In addition to initial training, the coaching and supporting of operators in the field is just as essential. Brian Cilinder-Hansen (CPH) says: “At the Security checkpoint in CPH on-the-job training by colleagues is an integral part of our education and daily training. Our experience shows that being trained by your peers create a safe space, where the employees trust each other and dare to ask questions. This is a great way to learn and an invaluable part of improving operator performance.” Even though operators have had sufficient training and gained the right skills for the job, knowing that someone is around to help when needed can be comforting.

For operators to feel satisfied in their job, it’s important that they’re comfortable and confident in their capabilities as a screener. Matching operators with the right training and coaching methods should increase the motivation to keep improving their performance and learning curve. This leads to the last subject, operator involvement & personal motivation.



Operator involvement and personal motivation

Brian Cilinder-Hansen-01.png
We can all have a bad day, if we experience problems at home or a colleague at work says something that affects us mentally.
— Brian Cilinder-Hansen Security Process Manager Copenhagen Airport (CPH)

Humans in general have an intrinsic need to belong. Responding to this need by involving operators in optimization and development processes, could contribute to higher operator satisfaction. Yet, the complexity of the manner in which operators are involved in these processes, can be quite delicate. Sofie Eyckerman (G4S) elaborates: “It’s very easy to ask them [operators] questions and think that they now feel involved, but if you don’t show them what you do with their answers and they don’t get feedback in some way, the effect can be the exact opposite.” A balance should be found herein, for operators to be satisfied about this involvement.

The last, but certainly not least, aspect that influences operator satisfaction according to the field experts is the personal motivation of an operator. Motivation can vary from day to day, and that’s okay. “We can all have a bad day, if we experience problems at home or a colleague at work says something that affects us mentally”, Brian Cilinder-Hansen (CPH) says. That is something that cannot be managed - it’s only human. What can be managed is a fitting shift planning with sufficient breaks and an open environment where operators feel like they have someone to talk to. An open atmosphere that inspires growth: operators that are willing to grow, that embrace their responsibilities and find joy in their work; those are the operators with the best odds to find satisfaction in their job.

Banner Blog Operator Satisfaction-01.png

We would like to thank Brian Cilinder-Hansen, Sofie Eyckerman and Maykel Bloom again for their openness and personal insight into their vision on operator satisfaction and their valuable contribution to our discovery of the link between operator satisfaction and operator performance.

Human Factor and Screener Performance assessments at Point FWD

Explore what Point FWD has to offer in terms of Human Factor and Screener Performance assessments.

Point FWD publishes 3-part knowledge article series on CT screening.

This paper is built around three main pillars deep diving into the operator training perspective when transitioning towards CT. equipment in security checkpoints.