There are different sensor technologies for detecting mines.  The simplest consist in metal detectors.  These sensors are simple, lightweight and easy to use.  However, they only detect mines that have metal parts, and they are inefficient with non-metallic mines (plastic mines).  Other sensor types are required in such cases, such as sensors based on ground-penetrating radar (GPR), chemical sensors or artificial noses.

An efficient detection system, it is commonly thought, should blend different technologies. The DYLEMA project, however, is devoted to the development of mobile-robotics techniques for landmine identification and location.  The project’s scope does not include any sensor development.  Therefore, the simplest course would be to select a metal detector as the demining sensor, just to help in the detection and location of potential alarms.  After a suspect object is detected, its location must be marked in the system database for further analysis and possible deactivation.

The Schiebel AN-19/2 commercial mine-detecting set is used for the DYLEMA project’s purposes.  This detector is in service in the US Army as well as in several NATO countries.  It has been designed to detect very small metallic objects, typically mines with a very small metal content.  This detector may be seen in Figure 1 and has been chosen as the mine detecting set for the DYLEMA project

.

The sensor head consists of a basic support (S) made of Arnite that holds the metal detector plus additional range sensors -infrared sensors (IR)- for detecting the ground and controlling sensor-head height and attitude.  These sensors are located in pairs defining the upper and lower limits of the band in which the sensor head works.  This array allows the controller to adapt the sensor head to terrain irregularities. Figure 2 shows the sensor head support and the position of the sensors.

The sensor head also needs additional elements for detecting objects in the way. For this purpose 12 flex sensors (plastic ribbons which change electrical resistance as they bend) have been provided to detect objects in the sensor head’s trajectory. The electric resistance of these sensors depends on the bend of the sensor, so they are analog devices; however, we use a threshold to configure an on/off device. Therefore, when a given sensor bends the controller is informed of the position of the collision in the sensor head reference frame. Figure 3 depicts the support including the metal detector and the flex sensors.

Figure 5. Detection areas with the Lateral Sensor (Top view of the sensor Head)

The version 1 of the sensor head presented the problem of the large number of flex sensors and thus the large number of I/O channels required in the system to detect objects in the way. This part of the sensor head is called Lateral Sensor.

An easier solution for the Lateral Sensor consists in using just a pair of flex sensors as shown in Figure 4. The lateral sensor is then built with two sliding circular platforms (inner and outer) using a non-metallic round, thin and hollowed mobile flat platform (outer) which is able to slide on top of a second fixed platform (inner). The two flex sensors are placed 90 degrees out of phase, as shown in Figure 4, so that by combining the signal from the flex sensor we can determine if a contact has occurred. In such a case a fuzzy logic acquisition system can determine what zone of the outer frame is in contact with either the ground or an object in the sensor head path.

The Lateral Sensor used finally 4 flex sensors (45 degrees out of phase) to avoid the problems derived from the relative motions of inner and outer platforms. Figure 5 shows the different detection areas.

Figure 6. Sensor head base on a network of distance sensors

    The third version of the sensor head relies on a network of non-contact, distance sensors. Each sensor is based on a pair of infrared emitter-receivers with a specified operational range of 4- 30 cm. The sensor network is divided into two groups: lateral-sensor group and ground-sensor group. The first group takes care of detecting objects in the sensor-head path. The second group is devoted to compute the average plane underneath the sensor head. This is required for both controlling the distance between sensor head and ground and accommodating the sensor head to terrain irregularities. The main advantage of this new sensor head relies on the no-contact nature of the device with the environment.