The focus of this research strand was macroscopic (cm scale) self-assembly (see Ref.  for an overview). All experiments were done by employing various instances of a water-based self-assembly modular robot called Tribolon and simulations thereof. Using these systems, we addressed two important issues of macroscopic self-assembly: 1) How does the morphology (size and shape) of the modules affect the yield of the self-assembly process (i.e. the final amount of the desired aggregate structure)? And 2) do inter-module connection mechanisms exist which are more suitable for water-based self-assembly? We start by describing our first study (for details see Ref. ) in which we studied wedge-shaped tiles of different size and shape by performing two types of experiments: (a) experiments with the real system and (b) computer simulations thereof. Given that our aim was also the identification of general principles of self-assembly, we hypothesized that by performing more abstract kind of computer simulations, such general principles could be uncovered or better understood. In , we presented a mathematical analysis of our modular system based on chemical rate equations. The results point to a power-law relationship between yield rate and shape of the tiles. Using the real system, we further demonstrated how through a single parameter (an externally applied electric potential) it was possible to control the self-assembly of propeller-like aggregates. The results (also described in ) seem to provide a starting point (a) for quantifying the effect of morphology on the yield rates of macroscopic self-assembly processes and (b) for assessing the level of modular autonomy and computational resources required for emergent functionality to arise. How well these results can be ported to lower scales (um or nm scale) is still subject to investigation. In the second study , we designed and realized a novel type of inter-module connection mechanism for waterborne (macroscopic) modular robotic systems. The proposed mechanism used the thermoelectric effect to cool down and freeze the water between two modules thus causing them to attach to each other. We validated the feasibility of this mechanism by embedding a Peltier heat pump (m = 0.8 g) in two types of cm scale self-assembly systems, one in which the modules were free to move and one in which the modules were linked to each other through hinges. Our experimental results demonstrate that the proposed Peltier-based connector has (a) a high bond strength/weight ratio for a rather large range of temperatures, and (b) is rather robust against misalignments between docking modules, making it a useful alternative to current connection mechanisms for small scale low autonomy self-assembly systems. To summarize, although it is still unclear how knowledge acquired at the cm scale can be transferred to lower scales, it seems to be the case that there is a lot to be learned about the general principles of self-assembly by employing prototypical macroscale models such as Tribolon (see Ref. ).
 Miyashita, S., Kessler, M. and Lungarella, M. (2008). How morphology affects self-assembly in a stochastic modular robot. In: Proc. of IEEE Int. Conf. on Robotics and Automation, pp. 3533-3538.
 Miyashita, S., Casanova, F., Lungarella, M. And Pfeifer, R. (2008). Proc. of IEEE Int. Conf. on Intelligent Robots and Systems. (to appear)
 Miyashita, S., Lungarella, M. and Pfeifer, R. (2009). Tribolon: water-based self-assembly robot. In: Artificial Models in Hardware. (to appear)