The most important anthropogenic sources of metals in aquatic ecosystem are certainly wastewaters, that are being discharged untreated or with different levels of purity, so they can cause many changes in the stream / recipient. Heavy metals pass go through biogeochemical process with different retention time in different parts of atmosphere. They do not decompose and have the ability of bioaccumulation, because they are being retained in plants, animals and nature in general. This is an extremely heterogeneous group of elements in terms of biological and ecological effects. Large numbers of metals is essential for proper functioning of the human body and appertain to the group of essential elements. A deficiency of heavy metals on one side can lead to serious symptoms; and on the other, their presence in slightly elevated concentrations can lead to serious illnesses. Their toxicity depends on concentration, and the allowed concentration range varies from metal to metal. Concentrations in which heavy metals can occur depend from source of pollution and features of system in which they are found, so they can range from traces to very high concentrations. Water, as well as air can receive large amounts of pollutants, but beside ability of self-cleaning, some pollutants among which are also heavy metals leads to modification of water quality to that level that it becomes useless for many purposes. From that reason most of European countries have issued maximum allowed amount of heavy metals in industrial waters that are being discharged into natural flow. Some of those values are given in table 1.

This paper presents a visual/motor behavior learning approach, based on neural networks. We propose Behavior Chain Model (BCM) in order to create a way of behavior learning. Our behavior-based system evolution task is a mobile robot detecting a target and driving/acting towards it. First, the mapping relations between the image feature domain of the object and the robot action domain are derived. Second, a multilayer neural network for offline learning of the mapping relations is used. This learning structure through neural network training process represents a connection between the visual perceptions and motor sequence of actions in order to grip a target. Last, using behavior learning through a noticed action chain, we can predict mobile robot behavior for a variety of similar tasks in similar environment. Prediction results suggest that the methodology is adequate and could be recognized as an idea for designing different mobile robot behaviour assistance.