论文：StanfordDoggo:AnOpen-Source,Quasi-Direct-DriveQuadruped

5星 · 超过95%的资源需积分: 35 154 浏览量更新于2023-03-03 评论 6 收藏 2.13MB PDF 举报

身份认证购VIP最低享 7 折!

领优惠券(最高得80元）

资源详情

资源评论

资源推荐

Stanford Doggo: An Open-Source, Quasi-Direct-Drive Quadruped

Nathan Kau, Aaron Schultz, Natalie Ferrante, Patrick Slade

Abstract— This paper presents Stanford Doggo, a quasi-

direct-drive quadruped capable of dynamic locomotion. This

robot matches or exceeds common performance metrics of state-

of-the-art legged robots. In terms of vertical jumping agility, a

measure of average vertical speed, Stanford Doggo matches the

best performing animal and surpasses the previous best robot

by 22%. An overall design architecture is presented with focus

on our quasi-direct-drive design methodology. The hardware

and software to replicate this robot is open-source, requires

only hand tools for manufacturing and assembly, and costs less

than $3000.

I. I

NTRODUCTION

Legged robots provide a highly mobile platform to traverse

difﬁcult terrain and are ideal for accomplishing tasks that

are repetitive, strenuous, or dangerous. Many state-of-the-

art legged robots have achieved remarkable feats including

high speed locomotion [1], [2], agile maneuvers [3], [4], [5],

and traversing difﬁcult terrain [6], [7], [8], [9]. Some designs

store energy, such as with a parallel-elastic leg mechanism, to

achieve dynamic motion and are not capable of continuously

agile motion required to accomplish many tasks [10], [11],

[12]. Only robots that carry their power supply and are

capable of repeated jumping or locomotion are considered

in this work. These platforms take years of development and

are frequently expensive, custom designs.

Many metrics characterize the performance of legged

robots: steady velocity during running, jump height, and

vertical jumping agility. Vertical jumping agility quantiﬁes

how quickly an animal can change its energetic state, approx-

imating the vertical climbing speed through a series of jumps

[5], [13]. This metric correlates to locomotion performance

as the distance a robot can jump increases its ability to

overcome obstacles, improving path-planning capabilities

[14]. While current quadruped robots are a popular platform

capable of carrying payloads, performing manipulation, and

fall recovery, they are unable to match the vertical jumping

agility of specialized monopod robots [5], [15] that attempt to

emulate the animal with the best jumping agility, the galago

(Galago senegalensis) [16].

Legged robots require a trade-off between energy efﬁ-

ciency to accomplish their task and sensitivity to safely

interact with their environment. Often, robots employ a high

reduction gear train to increase the effective torque produced

by the motor. This rigid gear train requires compliance

to be designed in series with the motor, referred to as a

series-elastic actuator [17]. Direct-drive (DD) robots do not

employ any speed reduction between the motor and output

The authors are with the Department of Mechanical Engineering, Stanford

University, Stanford, CA 94305 USA (e-mail: nathankau@stanford.edu).

shaft, allowing the sensitive motor to implement tunable

compliance through control at kHz timescales and maintain

a larger control bandwidth than the series-elastic actuators

[18]. A compromise is achieved with a quasi-direct-drive

(QDD) which uses a single stage reduction with a ratio less

than 10:1. This single stage increases torque output at the

expense of control bandwidth, but maintains the ability to

backdrive the motor which allows sensing of external forces

based on motor current [19].

This paper presents Stanford Doggo (Fig. 1), a QDD

quadruped that matches or exceeds common performance

metrics of state-of-the-art legged robots. The mechanical

design utilizes a belt drive as a lightweight QDD transmis-

sion to increase the effective torque while maintaining low

reﬂected inertia and transparency to enable sensitive control.

The QDD enables Stanford Doggo to match the performance

of the animal with the best vertical speciﬁc agility [16], a

22% improvement over the previous best robot [15]. Stanford

Doggo uses completely open-source hardware and software,

enabling full replication [20]. The total cost of materials

and machining costs is less than $3000 and requires only

hand tools for assembly. We hope to advance research and

education in the ﬁeld of legged robotics by lowering the cost

and resources required to have access to a state-of-the-art

robot.

Fig. 1: Stanford Doggo: an open-source, quasi-direct-drive

quadruped robot.

本内容试读结束，登录后可阅读更多

下载后可阅读完整内容，剩余6页未读，立即下载

夜阑雨

2020-02-02

7页，算是摘要，可以看看

ViatorSun

粉丝: 1w+
资源: 28

会员权益专享

论文：Stanford Doggo: An Open-Source, Quasi-Direct-Drive Quadruped

评论3

会员权益专享

最新资源

论文：Stanford Doggo: An Open-Source, Quasi-Direct-Drive Quadruped

评论3

斯坦福大学开源四足DogGo论文

开源项目-DATA-DOG-godog.zip

StanfordQuadruped

C:\Users\chenjing>java -mx4g -cp "*" -cp "D:\stanford\stanford-corenlp-4.5.4.jar"edu.stanford.nlp.pipeline.StanfordCoreNLPServer -port 9000 -timeout 15000 Unrecognized option: -port Error: Could not create the Java Virtual Machine. Error: A fatal exception has occurred. Program will exit.

C:\Users\chenjing>java -mx4g -cp "*" edu.stanford.nlp.pipeline.StanfordCoreNLPServer -port 9000 -timeout 15000 错误: 找不到或无法加载主类 edu.stanford.nlp.pipeline.StanfordCoreNLPServer

stanford doggo三维模型

/path/to/stanford-corenlp 需要替换成 stanford-corenlp 所在的实际路径，路径一直错误

stanford-corenlp-full-2018-10-05模型下载

stanford-corenlp-full-2015-12-09

斯坦福 stanford doggo 图纸

无法在终端或命令行中运行cd /path/to/stanford-corenlp java -mx4g -cp "*" edu.stanford.nlp.pipeline.StanfordCoreNLPServer -port 9000 -timeout 15000

介绍A*路径规划算法的外文文献

stanford-corenlp

TF-IDF java

如何安装StanfordParser

stanforddoggo

stanford corenlp

会员权益专享

最新资源