Autonomous robotic additive manufacturing through distributed model-free deep reinforcement learning in computational design environments

The objective of autonomous robotic additive manufacturing for construction in the architectural scale is currently being investigated in parts both within the research communities of computational design and robotic fabrication (CDRF) and deep reinforcement learning (DRL) in robotics. The presented study summarizes the relevant state of the art in both research areas and lays out how their respective accomplishments can be combined to achieve higher degrees of autonomy in robotic construction within the Architecture, Engineering and Construction (AEC) industry. A distributed control and communication infrastructure for agent training and task execution is presented, that leverages the potentials of combining tools, standards and algorithms of both fields. It is geared towards industrial CDRF applications. Using this framework, a robotic agent is trained to autonomously plan and build structures using two model-free DRL algorithms (TD3, SAC) in two case studies: robotic block stacking and sensor-adaptive 3D printing. The first case study serves to demonstrate the general applicability of computational design environments for DRL training and the comparative learning success of the utilized algorithms. Case study two highlights the benefit of our setup in terms of tool path planning, geometric state reconstruction, the incorporation of fabrication constraints and action evaluation as part of the training and execution process through parametric modeling routines. The study benefits from highly efficient geometry compression based on convolutional autoencoders (CAE) and signed distance fields (SDF), real-time physics simulation in CAD, industry-grade hardware control and distinct action complementation through geometric scripting. Most of the developed code is provided open source.

Supplementary Information

The online version contains supplementary material available at 10.1007/s41693-022-00069-0.

Genetic polymorphism of the alpha 2-adrenergic receptor is associated with increased platelet aggregation, baroreceptor sensitivity, and salt excretion in normotensive humans

It is likely that a number of independent heritable traits, each encoded by a singular gene, contribute to pathologic elevations in blood pressure in humans. Genetic polymorphisms of individual genes may result in intermediate phenotypes which, by themselves, do not raise blood pressure, but, coupled with environmental or epistatic forces, contribute to the prevalence of human hypertension. The gene for the alpha 2-adrenergic receptor encoded by chromosome 10 (C10 A2AR) is polymorphic, and Southern blotting with a cDNA probe following restriction enzyme digest of this gene results in fragments of either 6.3 kb or 6.7 kb in size. We reported an association between homozygosity for the 6.3 kb allele and hypertension in blacks. Blacks with hypertension also have an increased risk for thrombotic stroke, increased baroreceptor sensitivity, and decreased sodium excretion. We noted that the C10 A2AR, which modulates norepinephrine release in blood-pressure-regulating regions of the brain, is also expressed on platelets and in the kidney. We postulated that functional changes associated with the C10 A2AR gene polymorphism could be responsible for increased baroreceptor sensitivity, epinephrine-mediated platelet aggregation, and decreased sodium excretion in some individuals.(ABSTRACT TRUNCATED AT 250 WORDS)

Acute behavioural comparisons of toluene and ethanol in human subjects

A comparison of toluene and ethanol (EtOH) induced changes in central nervous system (CNS) function and symptoms were evaluated in two studies, and when possible the effects of toluene were expressed in EtOH equivalent units. The toluene concentrations were 0, 75, and 150 ppm, bracketing the American Conference of Governmental Industrial Hygienists threshold limit value (ACGIH TLV) of 100 ppm. The socially relevant EtOH doses were 0.00, 0.33, and 0.66 g EtOH/kg body weight, equivalent to two and four 3.5% 12 ounce beers. Forty two paid college students were used in each study. In the first study, subjects were exposed to toluene and an odour masking agent menthol (0.078 ppm) for seven hours over three days. In the second study EtOH or a placebo was administered at 1530 across three days also in the presence of menthol. Verbal and visual short term memory (Sternberg, digit span, Benton, pattern memory), perception (pattern recognition), psychomotor skill (simple reaction time, continuous performance, symbol-digit, hand-eye coordination, finger tapping, and critical tracking), manual dexterity (one hole), mood (profile on mood scales (POMS), fatigue (fatigue checklist), and verbal ability were evaluated at 0800, 1200, and 1600. Voluntary symptoms and observations of sleep were collected daily. A 3 x 3 latin square design evaluated solvent effects simultaneously controlling for learning and dose sequence. An analysis of variance and test for trend were performed on am-pm differences reflecting an eight hour workday and on pm scores for each solvent, in which subjects were their own control Intersubject variation in absorbance was monitored in breath. A 5 to 10% decrement was considered meaningful if consistent with a linear trend at p less than 0.05. At 150 ppm toluene, losses in performance were 6.0% for digit span, 12.1% for pattern recognition (latency), 5% for pattern memory (number correct), 6.5% for one hole, and 3% for critical tracking. The number of headaches and eye irritation also increased in a dose-response manner. The greatest effect was found for an increasing number of observations of sleep. A range of 2 to 7% decrements suggest the ACGIH TLV of 100 ppm toluene may be a good estimate of the biological threshold supporting a re-evaluation of the TLV. At 0.66 g EtOH/kg body weight symptoms and performance decrements were 6.6% for digit span, 9.2% for pattern recognition, 4.0% for continuous performance, 7.9% for symbol-digit, 16.5% for finger tapping, 6.2% for critical tracking, and 5.2% for the one hole test. The EtOH equivalents at 150 ppm toluene for digit span (0.56g EtOH/kg/body weight), the latency for pattern recognition (0.66 g EtOH kg body weight), and the one hole element "move" (0.37 g EtOH kg body weight) show that the first two measures would be affected at or above the 50 mg% blood alcohol concentration. This concentration is recognised as the lowest alcohol concentration associated with increased numbers of automobile accidents. The results suggest that EtOH may be a useful acute standard to compare the effects of various industrial solvents and support investigating an association between exposure to solvents and increased risk to safety in industry.