The intelligent sorting system has been widely applied in large-scale pig farming. Some functions can also be realized, including slaughter weight uniformity, the labor saving of manual grouping, and the minimum stress responses of the pigs. In this study, a systematic evaluation was implemented to explore the effects of the intelligent sorting system on both individual behavior and production performance of the fattening pigs. Conventional small-group feeding was also compared with the large-group feeding. Two independent but complementary experiments were carried out on the behavior and production. In the behavior experiment, two treatments were established: a control group with conventional small-group feeding (6 pigs per pen, 4 pens in total) and an experimental group with the intelligent sorting system (60 pigs per pen, 1 pen). Specifically, via the cameras pre-installed above the pens, 24-hour continuous behavioral videos of the pigs were recorded on days 3, 35, and 75 of the experiment. Continuous manual observation and video playback were combined to analyze the frequency, duration, and spatial distribution of the feeding, drinking, defecating, urinating, and fighting behaviors. The behavioral data were then processed into the average group per pig, thus allowing for the comparison over the experimental treatments. In the production performance experiment, the control group consisted of 7 pigs per pen in 8 pens, while the experimental group shared the 63 pigs in a single pen under the Intelligent Sorting System. The initial weight, final weight, and feed intake of the pigs were recorded in the experimental period. The average daily feed intake (ADFI), average daily gain (ADG), feed conversion ratio (FCR), and slaughter weight uniformity were calculated to evaluate the influence of the intelligent sorting system on the growth performance. In terms of the behavior, the feeding duration of the experimental group decreased by 45.1%, compared with the control group. As such, the high feeding efficiency was then obtained to substantially reduce the redundant feeding behaviors. Fighting behavior, a key indicator of social stress, was also greatly reduced: The fighting duration decreased by 75.1% and fighting frequency by 68.0%, compared with the control group. Aggressive competition was effectively mitigated for the group's social stability. At the same time, the proportion of defecation and urination on the slatted floors decreased by 16.4% and 24.9%, respectively, in the experimental group. The behavioral frequencies per square meter of the slatted floor increased by 7.1 and 13.9 events, respectively. The pigs had an excellent preference for the slatted floor excretion. The excretory behaviors were more concentrated per unit area. The low relative proportion of the slatted area highlighted the need for further optimization of the functional area allocation in pen design. In terms of production performance, the average daily feed intake was higher in the experimental group, whereas the feed conversion ratio decreased by 10.1%, compared with the control group. Furthermore, the slaughter weight uniformity was improved by 18.8 percentage points, indicating that the growth variability was reduced within the group. The individual growth efficiency greatly contributed to more uniform market-ready pigs. In conclusion, the intelligent sorting system improved both behavioral efficiency and production performance of the fattening pigs. The stress behaviors (such as fighting) were also reduced due to the high feeding efficiency and slaughter weight uniformity. The welfare-oriented precision was supported in pig farming practices. Modern technology can offer a feasible pathway toward sustainable, efficient, and animal-friendly production. Future studies should focus on the spatial layout of the functional areas in the system. Its effectiveness can be improved to validate the long-term impact on different housing and livestock management.

Lighting is one of the critical environmental indicators in the intensive pig production. There is the significant influence on the growth, reproductive performance, and behavioral patterns of the replacement gilts. This study aims to identify the optimal protocols of the artificial lighting. Four lighting treatments were also evaluated, including the L1 (white light, 40-50 lx), L2 (white light, 100-120 lx), L3 (green light, 100-120 lx), and L4 (white light with gradually varying intensity). Sixteen replacement gilts with an initial body weight of (86.00±9.50) kg were randomly assigned to four lighting treatments (4 pens per unit, 1 gilt per pen) for a 35-day trial. Body weight and feed consumption were measured on days 1 and 35. The data was then used to calculate the average daily gain (ADG), average daily feed intake (ADFI), and feed conversion ratio (F:G). Blood samples were collected to analyze serum hormone levels, including the gonadotropin-releasing hormone (GnRH), estradiol (E2), follicle-stimulating hormone (FSH), progesterone (PROG), cortisol, immunoglobulin A (IgA), luteinizing hormone (LH), and melatonin (MT). Furthermore, the behavioral observations were conducted using video recording systems. Continuous manual recordings were performed from 07:00 to 19:00 on days 7 and 14. The behaviors were also quantified, such as lying, standing, walking, feeding, drinking, defecating, and urinating. The results demonstrated that there was no difference in the growth performance (ADG, ADFI, or F:G) among the four lighting protocols (P > 0.05). However, some significant differences were observed in the physiological and behavioral parameters. L2 (100-120 lx white light) significantly enhanced the reproductive performance. Compared with the L1, the GnRH levels in the L2 increased by 15.74% (P< 0.05). The E2 levels in the L2 were 16.42% and 14.30% higher than those in the L3 and L4, respectively (P< 0.05). The PROG concentration in the L2 was 29.64% higher than that in the L1. While the FSH levels in the L2 exceeded those in the L4 by 44.98% (P< 0.05). The L2 enhanced the immune function, with the IgA concentration increasing by 14.72% and 23.44%, respectively (P< 0.05), compared with the L3 and L4. The L4 reduced the stress responses, with the cortisol concentrations being 21.96% and 18.84% lower than those in L2 and L3, respectively (P< 0.05). The L4 improved the behavioral functional zoning for the animal welfare. Compared with the L2 and L3, the L4 increased the lying rates on the solid floors by 86.88% and 71.77%, respectively. The urination rates on the slatted floors increased by 69.28%, compared with the L3 (P< 0.05). In conclusion, the lighting protocols were dominated the replacement gilt physiology and behavior. Growth performance remained unchanged aceoss all lighting treatments. White light at 100-120 lx (L2) also shared the most beneficial to the reproductive function. The gradually varying light intensity (L4) was minimized the stress responses to the behavioral organization. The precise production was achieved in the customized lighting strategies. For instance, the white light with 100-120 lx was improved the reproductive capacity of the replacement gilts, or the gradient white light was adopted to enhance the animal welfare levels. Considering the reproductive purpose of replacement filts, it is recommended to provide a 100-120 lx white light environment in replacemnt gilt house to improve their reproductive performance and immune function.This finding can provide the foundational framework for the precision light environment in the commercial pig farming, particularly for the efficiency and animal well-being.

Pigsty cleaning has been one of the most crucial working links in the treatment of manure and sewage in pig farms. The health status of pig herds can depend mainly on the cleaning performance. However, the current manual cleaning of pigsties cannot fully meet the large-scale production in recent years. It is also required for long-term exposure to higher humidity and concentrations of harmful gases inside pigsties, such as ammonia and hydrogen sulfide, compared with the outdoor environments. Moreover, manual manure cleaning has been the more unfavorable activity with an adverse impact on physical health, due mainly to the high labor cost and intensity. Some stress can also be likely to affect animal welfare during manual cleaning. Therefore, cleaning robots can serve as an effective solution to replace manual operations in the future. This study aims to review the research and application progress of manure-cleaning robots in the modern management of pig farms. Three perspectives were also set: the cleaning scene requirements and spatial operation of pig farms, the flushing of empty pig pens, the manure cleaning on the ground of pigsties with pigs present, and the cleaning of underground manure ditches. The key technologies of pigsty cleaning robots were also involved in multiple aspects, including application scene, chassis types, driving modes, position, and navigation, as well as path planning, manure cleaning mechanisms, and manure identification. The flushing robot of the pigsty empty pen was used to clean the floor and pen facilities of pigsties after each batch of pigs was removed under the "all in and all out" management mode of large-scale pig farms. The chassis type was then classified into the wheeled and tracked one. Among them, the wheeled chassis was often combined with mechanical arms and high-pressure water guns, predominantly suitable for flushing operations in the aisles of pigsties. The ground manure cleaning robot in pigsties with pigs was designed for the daily cleaning of residual manure on the floor inside the pen or on the aisle of the confinement stalls. Therefore, it was required for accurate manure identification and positioning, navigation, and path planning. Meanwhile, the effective manure cleaning included the scrape-shovel and manure collection, according to the materials of the slatted floors in pigsties. The underground manure ditch was characterized by its narrow space and dim lighting. Therefore, the cleaning robot was adopted to replace the manual cleaning and the frequently malfunctioning manure scraper. A tracked chassis was combined with the high-pressure water flushing and shovel operation. Residual manure was regularly removed in the manure ditch, in order to ensure a better air environment in the pigsty. Some suggestions were given to focus on the research of cleaning robots, according to the current development and deficiencies of Chinese pigsties. A special chassis can be developed with lightweight, excellent load-bearing and off-road capabilities suitable for pigsty cleaning scenarios; An independent environmental perception and operation can also be required in the path planning system; The durable sensors resistant to moisture and corrosion are suitable for the harsh environment of pigsties; A digital collaborative management and control platform can be established to realize the collaborative work of multiple intelligent agents, such as pigsty cleaning robots; and automated facilities are also complementary to cleaning robots. Some emphasis can then be put on the pen and pigsty cleaning robot in the concept of welfare and healthy breeding. The finding can also provide a strong reference to promote factory-based, unmanned, and welfare-oriented intelligent pig breeding.