Hierarchical assembly of self-replicating spacecraft using distributed mechanisms and actuation in digital materials

Abstract

A system of flexural, actuating, and semiconducting elements of part-types necessary to assemble actuated robotic systems. These parts are joined with a common interface, interlocking with neighboring parts to form a regular lattice structure. Primary considerations for the design of the part interfaces include ease of assembly and the ability to transfer mechanical loads and electronic signals to neighboring parts. The parts are designed to be assembled vertically so structures can he built incrementally one part at a time. They can be easily fabricated at a range of length-scales using a variety of two-dimensional manufacturing processes. These processes include, for example, stamping and laminating, which enable high-throughput production. The simple mechanical interfaces between parts also enable disassembly allowing for reconfigurability and reuse. The interlocking nature of these assemblies allows loads to be distributed through many parallel load-paths.

Claims

1. A discrete assembly system comprising: a plurality of parts belonging to a finite number of types including node part and strut part, wherein the node part comprises a two dimensional structure with a top and a bottom, with strut-connecting slots along both top and bottom for receiving rigid connections to strut parts, wherein the strut part comprises a two dimensional structure with a top and a bottom, with node-connecting slots along both top and bottom for receiving rigid connections to node parts, the plurality of parts constructed to be joined with a common interface; wherein the node parts additionally comprise at least two node-assembly slots offset from respective edges of the two-dimensional structure and disposed only on the top of the node part, the node-assembly slots having larger length than the strut-connecting slots; wherein the parts, when joined, form a regular lattice structure; and wherein, the parts are configured to be assembled vertically to make nodes solely by interconnection of node-assembly slots, and wherein assembled nodes may be interconnected through connection of node parts with strut parts at strut-connecting slots, allowing a structure to be assembled one part at a time.

2. The system of claim 1 wherein the parts are fabricated at different length scales by a 2-dimensional manufacturing processes.

3. The system of claim 1 wherein the parts interlock, allowing loads to be distributed through a plurality of parallel load paths.

4. The system of claim 1 wherein nodes are assembled from four identical node parts, each with four slots, each slot constructed to rigidly connect to a strut part.

5. The system of claim 4 wherein struts are constructed from strut parts.

6. The system of claim 5 wherein a quasi-isotropic structure is assembled from a plurality of nodes and struts built in planes that can be linked to one-another.

7. The system of claim 6 wherein the struts change orientation offset by every two nodes.

8. The system of claim 5 wherein struts can be flexural struts that can be constructed from two-dimensional laminate parts or from folded laminate parts.

9. The system of claim 5 wherein struts can be formed from two-part types, one of high-permeability material and the other a mandrel upon which a coil is wound.

10. The system of claim 9 wherein the high permeability material is iron or steel which houses one or more magnets to form an E-coil.

11. The system of claim 2 wherein the structure is assembled in a millimeter, sub-millimeter or micro-scale.

12. The system of claim 8 wherein the flexural struts are constructed to have one, two or three degrees of freedom.

13. The system of claim 12 wherein the flexural struts are manufactured using with individual laminate manufactured layers individually machined and bonded together.

14. The system of claim 13 wherein the flexural struts are bonded together using flexible double-sided tape.

15. The system of claim 12 wherein a revolute joint is constructed using two 1-degree of freedom struts with two 2-degree of freedom struts.

16. The system of claim 15 wherein the revolute joint has two orthogonal revolute degrees of freedom.

17. A discrete assembly system comprising: a plurality of parts belonging to a finite number of types including node part and strut part, wherein the node part comprises a two dimensional structure with a top and bottom, with strut-connecting slots along both top and bottom for receiving rigid connections to strut parts, wherein the strut part comprises a two dimensional structure with a top and bottom, with slots along both top and bottom for receiving rigid connections to node parts, the plurality of parts constructed to be joined with a common interface; wherein the node parts additionally comprise at least two node-assembly slots offset from respective edges of the two-dimensional structure and disposed only on the top of the node part, the node-assembly slots having larger length than the strut-connecting slots; wherein the parts, when joined, form a regular lattice structure; wherein, the plurality of parts are configured to be assembled vertically to make nodes solely by interconnection of node-assembly slots, and wherein assembled nodes may be interconnected through connection of node parts with strut parts at strut-connecting slots, allowing a structure to be assembled one part at a time; wherein the plurality of parts are fabricated at different length scales by 2-dimensional manufacturing processes; wherein the plurality of parts interlock, allowing loads to be distributed through a plurality of parallel load paths; and wherein the finite number of parts types comprises two parts types.

Description

DESCRIPTION OF THE FIGURES

(1) Attention is now directed to several figures that illustrate features of the present invention.

(2) FIG. 1 shows an example geometry made from parts.

(3) FIG. 2 shows design and modeling of actuator parts.

(4) FIG. 3 shows mechanisms assembled from discrete parts.

(5) FIG. 4 shows experimenting with circuit and logic routing.

(6) FIG. 5 shows an example of struts and nodes and of an assembly of both.

(7) FIG. 6 shows another example of a structure.

(8) FIG. 7 and FIG. 8 show actual laboratory structures constructed.

(9) FIG. 9 shows how a regular rigid structure can be fulfilled using struts composed of two identical two-dimensional parts.

(10) FIGS. 10-12 depict different length scales showing the possibility of 3D-printing, laser-cutting, and Wire-EDM'ing the parts.

(11) FIG. 13 is a walking motor shown in three different scales.

(12) FIG. 14 shows a single degree-of-freedom part (left) and a two degree-of-freedom part (right).

(13) FIG. 15 shows a series of experiments that were performed to determine a tape/material combination with adequate thickness, ultimately settling on using 0.0125″ delrin with 6miI 3M 410M double-sided tape.

(14) FIG. 16 shows parts being picked off with tweezers.

(15) FIG. 17 shows a simple jig that can be used to align the individually machined layers and press them together using an acrylic roller.

(16) FIG. 18 shows that the alignment between layers does not have to be not perfect.

(17) FIGS. 19-22 show that, once fabricated, these parts can be assembled to form a variety linkages and mechanisms.

(18) FIG. 23 shows a “leg” with two orthogonal revolute degrees of freedom.

(19) FIG. 24 shows a design example of an actuator.

(20) FIGS. 25-26 show such a voice coil actuator.

(21) FIGS. 27-29 show actuators are tested qualitatively by building them into mechanisms and powering them using an H-Bridge to switch current through the coils bi-directionally.

(22) FIG. 30 shows such a voice coil actuator.

(23) FIG. 31 shows a revolute joint that uses two 2-DoF flexural struts arranged in parallel to form a linkage which approximates straight-line motion.

(24) FIG. 32 shows an embodiment of a walking motor.

(25) FIG. 33 shows the small set of parts needed to assemble a walking motor.

(26) FIG. 34 shows a walking motor powered by two orthogonally oriented voice coil actuators.

(27) FIGS. 35-37 show the development and prototyping of this motor evolving from a manually driven test of the mechanism, to an actuated test, and finally showing its integration ruled slider to enable long-range motion.

(28) FIG. 38 shows a working prototype of a walking motor next to the same geometry assembled from millimeter scale parts.

(29) FIG. 39 shows three part types used to fabricate various cells.

(30) FIG. 40 shows various routing cells.

(31) FIG. 41 shows traces routed in 2-D.

(32) FIG. 42 shows split-conductivity struts.

(33) FIG. 43 shows that traces on the upper and lower portion of nodes can be isolated to allow crossover.

(34) FIG. 44 shows an additional crossover part.

(35) FIG. 45 shows a fabrication process.

(36) FIG. 46 shows continuity testing.

(37) FIG. 47 shows a Neopixel part.

(38) FIG. 48 shows wiring of Neopixels with crossover parts.

(39) FIG. 49-FIG. 51 show operation of the Neopixels.

(40) Several figures have been presented to aid in understanding the present invention. The scope of the present invention is not limited to what is shown in the figures.

DESCRIPTION OF THE PREFERRED EMBODIMENTS

(41) The present invention relates to parts are joined with a common interface, interlocking with neighboring parts to form a regular lattice structure. The parts are designed to be assembled vertically so structures can he built incrementally one part at a time. They can be easily fabricated at a range of length-scales using a variety of two-dimensional manufacturing processes. The simple mechanical inter-faces between parts also enable disassembly, allowing for reconfigurability and reuse. Furthermore, the interlocking nature of these assemblies allows loads to be distributed through many parallel load-paths and means that the size of the assemblies is not limited by the strength of the individual parts.

(42) Middle Out Machines

(43) Middle Out Machines are machines assembled from standardized functional building blocks. The present description describes types of middle out machines and is divided into four main sections: Geometry, Mechanism, Actuation, and Circuitry.

(44) The geometry section includes images and documentation of experiments with the physical geometry of the building block parts, for example FIG. 1 shows an example geometry made from parts. The actuation section includes design, modeling, and physical testing of actuator parts such as those in FIG. 2. The mechanisms section contains design, fabrication, and testing of mechanisms assembled from discrete parts such as those shown in FIG. 3. The circuits section contains experimental studies of circuit and logic routing in this building block architecture such as those shown in FIG. 4.

(45) Geometry

(46) There are two fundamental types of building blocks in this assembly architecture: node parts and strut parts. The building blocks are arranged hierarchically. Fabricated two-dimensional parts (node parts and strut parts) are first assembled into struts and nodes. Each node is assembled from four identical parts, each with four slots, which are used to rigidly connect to struts. The struts house most of the functionality of this building system and so can be built in a few different ways depending on their intended purpose. FIG. 5 shows an example of struts and nodes and of an assembly of both. FIG. 6 shows another example of a structure, while FIGS. 7 and 8 show actual laboratory structures constructed. These structures are built in planes and can be linked to assemble quasi-isotropic structures using struts and change orientation and vertical spaces that offset every two nodes.

(47) Regular rigid structure can be fulfilled using struts composed of two identical two-dimensional parts. Flexural struts can be assembled either from a combination of two-dimensional laminate parts, or from folded laminate parts. Furthermore, actuated struts can be assembled from two special part-types: one made of high permeability material such as iron or steel which houses two magnets and forms an E-core, and the other which serves as a mandrel on which a coil is wound. This is shown in FIG. 9.

(48) Scaling

(49) The flexibility of this building system enables fabrication at a variety of length-scales using a variety of processes. Shown in FIGS. 10-12 are a few different length scales showing the possibility of 3D-printing, laser-cutting, and Wire-EDM. For perspective, three scales of the node part spanning an order of magnitude from largest to smallest are shown. Initially there has been prototyping with centimeter scale parts, eventually, the assembly of millimeter, and sub-millimeter, parts will be targeted. The millimeter parts are made using the Wire-EDM with the 0.004″ wire, which cuts with about a 0.007″ kerf. Below this scale, parts can be made using the Oxford micromachining Laser which has a kerf of about 20 um. These parts are 75 um thick and are 1 mm long in their longest dimension.

(50) Pictured in FIG. 13 is a walking motor (discussed later in Actuation), assembled at the centimeter scale. A millimeter scale model is shown next to it assembled from rigid parts. Finally, a microscale part (‘˜’1 mm in longest dimension) is pictured on a nickel for scale.

(51) Mechanism

(52) Fabrication

(53) Mechanisms can be assembled by using parts with embedded flexural joints. Four part-types are useful for assembling a wide variety of mechanisms; these include rigid parts, and flexible parts with one, two, and three degrees-of-freedom. Shown in FIG. 14 are a single degree-of-freedom part (left) and a two degree-of-freedom part (right). The flexible hinges in these parts are close mirrors to conventional hinges at the macro scale. These parts can be efficiently manufactured using a laminate construction technique. Individual layers are laser machined and subsequently bonded together using a flexible double-sided tape.

(54) In order for these parts to rigidly connect with other parts in this construction system, the thickness of the laminate needs to match the standard part-thickness. A series of experiments were performed to determine a tape/material combination with adequate thickness, ultimately settling on using 0.0125″ delrin with '˜6miI 3M 410M double-sided tape. This is shown in FIG. 15.

(55) When laser cutting the individual layers, tabs are used to keep parts attached to a scaffold structure. Unfortunately this results in many little parts that need to be picked off with tweezers as shown in FIG. 16. The double-sided tape is also machined on the laser. With careful control of the laser speed and power, it's possible to cut through just the tape but not through the backing. A simple jig shown in FIG. 17 can be used to align the individually machined layers and press it together using an acrylic roller. The alignment between layers does not have to be not perfect, but rather sufficient to make a solid connection with neighboring parts as shown in FIG. 18.

(56) Parts with embedded hinges are fabricated using a laminate process. Layers are individually machined and then laminated together to form a composite. Flexural hinges are created by an absence of rigid material on either side of a more flexible core. The layers used to make the flexible part types are shown below.

(57) Assembly

(58) Once fabricated, these parts can be assembled to form a variety linkages and mechanisms as shown in FIGS. 19-22. A revolute joint can be made, for example, using two single degree-of-freedom struts along with two 2-degree-of-freedom struts, which are used to constrain the motion of an actuator. FIG. 23 shows a “leg” with two orthogonal revolute degrees of freedom.

(59) Actuation

(60) Design

(61) FIG. 24 shows a design example of an actuator.

(62) Simulation

(63) Voice coil actuators can be roughly modeled using magnetic circuit theory to calculate the magnetic field across the coil and calculate the resulting Lorentz force. From this it is determined that the initial design is expected to produce about 200 mN of force with 1 A of current. This assumes magnets that are 3 mm×2 mm×1 mm, a coil that is 1 mm thick, and with 0.25 mm of air-gap left on either side of the coil. FIGS. 25-26 show such a voice coil actuator. By increasing the magnet size to 6 mm×3 mm×2 mm, and the coil-thickness to 2 mm, the expected force can be increased to approximately 1N. This requires slightly more complicated geometry to achieve but may be do-able without introducing too many custom or difficult-to machine parts.

(64) Testing

(65) The actuators are tested qualitatively by building them into mechanisms and powering them using an H-Bridge to switch current through the coils bi-directionally. This is shown in FIGS. 27-29. The force produced by the actuators is measured using a table-top instron.

(66) Initial testing indicates that the actuator is only producing approximately 50 mN of force with 1 A of current (supplied directly from a current-limited power supply). This is corroborated by the fact that it can lift a 5 g bolt but fails to lift a 10 g mass under these conditions. While testing the coil at 1.5 A, the coil was becoming hot enough to melt the delrin mandrel parts (but not before destroying the parts). It took less than 30 seconds to reach that temperature (which means that 1.5 A is a firm upper-limit unless it is used with very short duty-cycles).

(67) Voice Coils

(68) Actuation in the structure is provided by individual voice coil actuators, which, like the rest of the structure, are assembled from a small set of parts. The actuators are composed of two main parts: a coil part and a magnet part. The coil part is itself composed of two struts, which form a mandrel around which, a coil is wound. The magnet part is composed of a permeable flux guide (made of steel or iron) in which two magnets are arranged such that their poles face each other. This produces a relatively large magnetic flux density between each of the magnets and the center core of the flux guide. The top and bottom of the coil fit in this gap with high magnetic flux and are free to slide in and out. When current is passed through the coil, it creates its own magnetic field which opposes that of the permanent magnets and effectively pulls or pushes the coil part relative to the magnet part. FIG. 30 shows such a voice coil actuator.

(69) The actuator needs to have its motion properly constrained in order to function. This can be accomplished in a variety of ways. FIG. 31 shows a revolute joint that uses two 2-DoF flexural struts arranged in parallel to form a linkage that approximates straight-line motion. This effectively constrains the actuator to move along that linear dimension. The actuator then transmits its force through a pair of 1 DoE flexural struts to convert its linear motion into revolute motion of part of the structure

(70) Walking Motors

(71) FIG. 32 shows an embodiment of a walking motor. In order to enable locomotion beyond the limited stroke of an individual actuator, a walking motor can be assembled which produces long-range motion through small repeated steps. This device is assembled from a small-set of standardized parts as shown in FIG. 33. This includes three node part-types (conducting, insulating, and split-conductive), three strut part-types (conducting, insulating, and 2-DoE flexural), and two actuator part-types (coil and magnetic).

(72) The walking motor is powered by two orthogonally oriented voice coil actuators as shown in FIG. 34. The stroke of one actuator, the “drive” actuator, is arranged to be parallel to the direction of long-range travel. The other actuator, or “latching” actuator, is used to engage and disengage the motor tip with a ruled slider. These actuations are transferred to the motor “tip” through a number of double-hinge struts that are arranged to enable independent vertical and horizontal shearing motions. For initial testing a ruled slider also shown in FIG. 34 is used to demonstrate how a series of small steps can create long-range motion. The slider has 1 mm wide notches with a pitch between notches of 2 mm.

(73) FIGS. 35-37 show the development and prototyping of this motor evolving from a manually driven test of the mechanism, to an actuated test, and finally showing its integration ruled slider to enable long-range motion.

(74) In the experiment, the voice coil actuators are driven with conventional off-board electronics; However, logic and electronics can be integrated within the same assembly architecture, enabling the assembly of the sequencing of the drive signals.

(75) Each voice coil actuator is driven by an H-Bridge motor driver and controlled by an ATmega32u4 microcontroller. In this experiment, the voice coils are driven with sinusoidal waveforms, one phase-shifted by 90 degrees. With a 800 mA main current supply, repeatable motion can be obtained. Beyond 1 A main current supply, the actuators can only be used for a limited time before their temperature rises to potentially damaging levels.

(76) The first working prototype was constructed from centimeter-scale parts. Millimeter (and microscale) parts can also be assembled. Pictured in FIG. 38 is a working prototype next to the same geometry assembled from millimeter scale parts.

(77) Circuitry

(78) With three part types shown in FIG. 39, various routing cells can be assembled as shown in FIG. 40. Then, through arrangement of the cells, traces can be routed in 2-D as shown in FIG. 41. With the addition of split-conductivity struts shown in FIG. 42, traces on the upper and lower portion of nodes can be isolated and allow crossover within the same node/cell as shown in FIG. 43. An additional crossover part shown in FIG. 44 is necessary to be able to route the bottom of one node to the top of a neighbor node (and vice versa).

(79) Fabrication

(80) These parts can be fabricated fairly easily by milling them from copper clad board. One embodiment uses 1/32″ thick copper clad FR1 board with 1 oz (1.4 mil) double-sided copper. The Roland MDX-540 is an ideal tool for this job because the automatic tool-changer enables the use of rough and finish tools to deal with rapid tool wear. To do double sided milling, two registration holes are drilled through a board and the fixture plate. Then ⅛″ taper pins are used to position the board with high precision after flipping it over. FIG. 45 shows a fabrication process.

(81) Testing

(82) The continuity of the conductive joints was tested by constructing a simple LED assembly. Ground is supplied by the bottom half of the node, and power is supplied by the top half as shown in FIG. 46. Many LED's can be wired in parallel using this routing scheme.

(83) A more complex routing example is needed to route traces for Neopixel parts. An example of a circuit routing that can done with this assembly architecture involves the wiring of an array of Neopixel smart RGB LED's. Each neopixel has four connections: power, ground, data in, and data out as shown in FIG. 47. The power and ground are wired in parallel, while the data in and data out allows the neopixels to be chained serially.

(84) To route this example, the split-conductive node and split-conductive strut are needed. Additionally, a crossover strut is required in order to have a signal change layer in the structure as shown in FIG. 48. Here, signal wires that control three neopixels independently are routed through the structure. Power, ground, and a signal data input signal are supplied externally at one node of the structure, and the signals are routed based on the construction of the structure. FIGS. 49-51 show testing of the assembled circuit. The RGB LED's cycle through their colors one at a time.

(85) Examples and illustrations have been presented that show features of the present invention and aid in understanding it. One skilled in the art will realize that numerous changes and variations are possible without departing from the spirit of the invention. Each of these changes and variations is within the scope of the present invention.