Autonomous mobile robot navigation systems are based on three principal kinds of techniques: map-based navigation, map-building-based navigation and mapless navigation. We propose a mapless method for trajectory description in unknown indoor environments. The method uses distance measurements from a 2D laser range finder, digitises the robot's visibility area, eliminates superfluous data and reorients their presentation with laws similar to those used in cellular automata. The landmarks are extracted and organised in a panoramic description called fresco. The frescoes which are validated by means of neighbourhood rules. The most informative frescoes are detected by means of two criteria and stored. The stored frescoes are considered as a human-like descritption of the robot's route and could be used by the robot to retrieve its route to its starting point.

Two groups of 15 multiparous cows in mid-lactation were used in a Latin square design experiment with 4-week experimental periods. The genetic milk protein concentration level was high in the first group and low in the second. Each group of cows was given in a random order three feeding levels that covered 85, 100 and 115% of energy requirements and 90, 110 and 125% of nitrogen requirements, respectively. In both groups, increasing level of feeding induced a significant increase in milk yield (+2·4 kg/d between lowest and highest levels) and in protein concentration (+1·7 g/kg). The proportion of paracasein in total proteins was not altered by either genetics or nutrition. The proportion of casein in total proteins was slightly increased by 0·5 percentage points (P<0·05) with the intermediate level of feeding. Plasmin and plasminogen activities were not significantly modified by the genetic milk concentration level. Plasmin activity significantly increased with nutrient supplementation, but only in animals of low genetic potential (+21% between low and high levels, P<0·01). Casein composition was not significantly altered by the genetics or level of nutrition. Over the whole range of individual measurements taken (n = 90), the relationships between casein or paracasein and total protein concentrations were linear and very narrow (R2 = 0·92 and 0·95, respectively). The proportion of casein or paracasein in total proteins significantly decreased as plasmin activity increased.

The composition and sensory characteristics of matured cheeses are controlled by a number of factors, among which the type of feed is important. The influence of feeding can be reflected by the presence in cheeses of terpenes and sesquiterpenes, compounds typically indicating their vegetable origin (Mariaca et al. 1997). Indeed, several investigators have already established that these compounds could characterize the forage even to a specific geographical location. Dumont & Adda (1978), Dumont et al. (1981), Guichard et al. (1987), Bosset et al. (1994) and Moio et al. (1996) were thus able to distinguish cheeses from lowland and upland regions. In most of these studies, the conditions under which milk was produced and processed were not completely controlled. Furthermore, all these studies analysed only the volatile fraction of cheeses and did not examine the volatile compounds in the forages used. Recently Bosset et al. (1994), in one part of the project described by Jeangros et al. (1997), showed that highland grass with a highly diversified botanical composition produces milk and cheeses with significantly different chemical compositions from those from lowland grass. To improve our understanding of the relationship between animal feed and cheese composition, we have investigated under controlled experimental conditions both the composition of the terpene and sesquiterpene fractions of four forages with different botanical diversities and the influence of those forages on the terpene fraction of cheeses.

Two groups of eight multiparous cows with different calving periods (November or February) were managed in the same way during lactation. During four 4 week experimental periods distributed over 12 months, the cows were fed on a diet composed of hay and concentrate (70[ratio ]30) in a restricted and controlled amount which varied according to their lactation stage, so as to cover the animals' requirements correctly. The animals' average lactation stage varied according to period from 26 to 298 d. Milk from each group was processed on two occasions during each period to make Saint-Nectaire-type cheese; the cheesemaking conditions were the same throughout. Lactation stage had an important effect on milk fat, protein and calcium contents but not on the casein[ratio ]protein ratio or phosphorus content. The milk pH and the urea content were higher in late lactation. The calcium concentration of milk was higher in late lactation but the soluble fraction was higher in early lactation. Despite higher protein contents, the maximal firmness of the coagulum of late-lactation milk was not different from that of early or mid-lactation milks. pH was higher in cheeses from late-lactation milks compared with those from early and mid-lactation milks. Cheeses from early lactation milks were more yellow than the others and had a lower dry matter fat content. In sensory analysis the odour of cheeses from early and late-lactation milks was less pleasant than that of those from mid-lactation milks. Cheeses from late-lactation milks were more melting and less firm than those from early or mid-lactation milks. Their taste was more intense and more persistent. At tasting, they were less appreciated than the others. These differences were linked to increased proteolysis in the cheeses made with late-lactation milks.

Forty-two multiparous dairy cows of three different breeds (Holstein, Montbéliarde and Tarentaise) were fed on the same type of forage (natural grassland) preserved in the form of either hay (H) or silage (GS), according to a changeover design (two 4 week periods). The proportion of concentrate in the diet and the energy and nitrogen contents were similar in both treatments. The milk produced by these cows was used for the manufacture of Saint-Nectaire type cheeses, under controlled and identical cheesemaking technological conditions. More cheese was produced with the H treatment milk. The cheeses made with the GS treatment milk were more yellow and tended to be more bitter. The other chemical and sensory characteristics did not differ much between the two treatments. Of the 51 volatile compounds identified, four were in significantly higher proportion in the GS than in the H cheeses. Cheeses produced from Tarentaise cows' milk were more yellow and their pH was higher than those made with the milk of Holstein or Montbéliarde cows. The cheeses from Montbéliarde and Tarentaise cows' milk were firmer, more melting and tastier than those made with the milk of Holstein cows. Although some trends were apparent, there were no significant differences in cheese volatile compounds for different breeds.