Lab on a Chip PAPER Cite this: Lab Chip, 2014, 14, 189 Received 23rd September 2013, Accepted 25th October 2013 DOI: 10.1039/c3lc51083b www.rsc.org/loc Control of soft machines using actuators operated by a Braille display† Bobak Mosadegh, ab Aaron D. Mazzeo, a Robert F. Shepherd, ab Stephen A. Morin, ac Unmukt Gupta, a Idin Zhalehdoust Sani, ab David Lai, c Shuichi Takayama cd and George M. Whitesides * ab One strategy for actuating soft machines (e.g., tentacles, grippers, and simple walkers) uses pneumatic inflation of networks of small channels in an elastomeric material. Although the management of a few pneumatic inputs and valves to control pressurized gas is straightforward, the fabrication and operation of manifolds containing many (>50) independent valves is an unsolved problem. Complex pneumatic manifolds—often built for a single purpose—are not easily reconfigured to accommodate the specific inputs (i.e., multiplexing of many fluids, ranges of pressures, and changes in flow rates) required by pneumatic systems. This paper describes a pneumatic manifold comprising a computer-controlled Braille display and a micropneumatic device. The Braille display provides a compact array of 64 piezoelectric actuators that actively close and open elastomeric valves of a micropneumatic device to route pressur- ized gas within the manifold. The positioning and geometries of the valves and channels in the micropneumatic device dictate the functionality of the pneumatic manifold, and the use of multi-layer soft lithography permits the fabrication of networks in a wide range of configurations with many possible functions. Simply exchanging micropneumatic devices of different designs enables rapid reconfiguration of the pneumatic manifold. As a proof of principle, a pneumatic manifold controlled a soft machine containing 32 independent actuators to move a ball above a flat surface. Introduction One emerging class of soft machines comprises tools fabri- cated by molding pneumatic channels or features into elasto- meric polymers; 1–4 these channels provide desired actuation upon pressurization and inflation. Pneumatic actuation using air as a working fluid has many advantages in the operation of soft machines: pressurized air has low viscosity, low mass, high availability, no environmental impact, and little cost. In order to automate actuation using pneumatic technologies, computer-controlled valves are usually used to regulate the delivery of pressurized gas, and the functions provided by the valves dictate, in turn, the level of control possible for the resulting machines. This paper describes a reconfigurable manifold useful for controlling pneumatically actuated soft machines; 1–7 it is composed of a computer-controlled Braille display, which provides a compact array of piezoelectric actuators, and an interchangeable micropneumatic device, which dictates the routing of pressurized fluids between inputs, outputs, and elastomeric valves. This design is based on an analogous use of a Braille display by Takayama et al. to control aqueous flows in microfluidic systems. 8 Microfluidic devices can be designed to actuate elasto- meric valves pneumatically to perform on-chip pumping and routing of fluids. 6,7,9–12 Because they are reliable, and can accommodate a wide range of pressures and flow rates, banks of computer-controlled solenoid valves are the most commonly used controllers for multi-channel pneu- matic systems. 2,3,6 With this approach, reconfiguring inter- connections among valves to enable different functions is difficult. This difficulty becomes progressively greater as the number of solenoid valves increases. There are several methods to reduce the number of exter- nal solenoid valves needed to control a large number of pneumatic outputs, and these methods can be classified into three categories: i) parallel instruction, ii) serial instruction, and iii) embedded instruction. 13 Serial instruction and a Department of Chemistry and Chemical Biology, Harvard University, 12 Oxford Street, Cambridge, MA 02138, USA. E-mail: gwhitesides@gmwgroup.harvard.edu b Wyss Institute for Biologically Inspired Engineering, Harvard University, 60 Oxford Street, Cambridge, MA 02138, USA c Department of Biomedical Engineering, University of Michigan, 2200 Bonisteel Blvd, Ann Arbor, Michigan 48109-2099, USA d Macromolecular Science and Engineering Center, University of Michigan, 2300 Hayward St., Ann Arbor, Michigan 48109, USA † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c3lc51083b Lab Chip, 2014, 14, 189–199 | 189 This journal is © The Royal Society of Chemistry 2014