Robotic Taxi Project

Let's Make Robots Projects Showcase
1

Robotic Taxi Project

Requirements

Build a robot to autonomously transport a single person.

Use case scenario.

A person wants to be taxied from a specified pickup point to their home.

The person would be able to reserve the robot via the internet by entering a pickup time and specify pickup and drop-off points. The website would determine robot availability and/or propose other time slots available for pickup. Reservations would be made via credit card (PayPal) on the website and payment for service would be charged at time of drop-off.

Business Requirements

1 ord). Must meet city ordinance for electric vehicle use.

“Ordinance requiring the electric vehicles to be licensed. Any driver of the vehicles must hold a valid driver license. The electric vehicles only would be allowed on roads with maximum speeds of 35 miles per hour or less.”

Technical requirements

1 ord).

a) Routing information will be entered into the system limiting routes to roads posted 35 miles per hour or less.
b) Vehicle speed, location, along with webcam video/audio feed of the vehicle path and passenger, in addition to other sensor data, will be monitored by dispatchers (who have valid drivers licenses). They can assist in vehicle operation by taking complete remote control of the vehicle at anytime.
c) The vehicle will fail safe. If the vehicle loses communication for more than one second flashers will activate and the vehicle will automatically come to a stop.
d) The passenger will have a red shutdown button that will stop the vehicle at anytime, activate flashers, and notify dispatch.

Communication

The robot will be controlled and monitored via bluetooth comm port during initial prototyping. The bluetooth signal will be extended via a bluetooth enabled cell phone connected to a 3G wireless data network in the final implementation.

Motor requirements
Must reach speeds of at least 20 MPH.
Run on two motors (for cost savings).

**Range **

Requirement
14 miles on a single battery charge

Design considerations (not requirements)
Mileage could be extended by adding:
gas powered generator, placement should be considered during the design phase.
Quick and easy battery replacement
Solar charging
Clearance
Should be able to clear a curb.

Winter Operation (will not be a requirement but is a consideration)
Snow (clearance, traction control), ice (traction control), heater (for occupants).

Weight
Platform must support at least 225 pounds.

Sensors

One forward facing camera will have a convex mirror in view (for passenger observation). A servo operated camera might make it into the final design for 360 degree viewing (no design is complete without ssc32 inclusion in my opinion). Bandwidth and cost are determining factors for limiting design to single camera use (as opposed to a multiple camera setup).

GPS, compass, accelerometers, gyros, proximity sensors.

Others are Welcome
Join in* by creating miniature taxies using lynxmotion motor controllers (and platform of your own design) to be considered for production franchise use.

I will abstract the motor controller communication layer in the software so I can write drivers to operate other controllers that support serial commands. You must use a motor controller that supports serial commands.
lynxmotion.com/Category.aspx?CategoryID=10

  • If you do add your miniature taxi into the on-line motor pool I cannot be held responsible for any robot that leaves a home pursuing active taxi service duty.

Four platform designs under consideration (in order of preference).

Round platform
One wheel on the right, one on the left, and a caster wheel in front and back.
This would have the most maneuverability.

Square platform - Three wheels
Two wheels in front with one caster wheel in back.

Square platform - Four wheels
Would use tank style steering.
Could run chain drive inside the chassis to get four wheel drive
Could get two more motors for four wheel independent drive for winter operation. Would also require additional motor controller.

Square platform - Four wheels with servo steering (like a car)
Probably too complex but could be applied to existing automobile platforms

The two wheeled platform could ultimately support this design.
gizmodo.com/5252544/i-want-robot-taxis
http://cache.gawker.com/assets/images/gizmodo/2009/05/robo-taxi_1_suk4Q_69.jpg

Timeline
Working prototype capable of carrying a single passenger before the last leaf falls in the year of 2009.

Shopping list

Motor Controllers
lynxmotion.com/Product.aspx?productID=623&CategoryID=10
robotshop.us/controleur-robot-ax2550-6.html
robotpower.com/products/sidewinder_info.html
ifirobotics.com/frc-robot-control-system-overview.shtml

Motors
npcrobotics.com/bbiq/kit.asp?prod=67&cat=26

S28-400
battlekits.com/robot_motors.htm

Probably not powerful enough but shows 3 wheel platform design
trossenrobotics.com/store/p/5761-Motor-Mount-Wheel-Kit-with-Position-Controller.aspx

Tire
NPC-PT5306
npcrobotics.com/products/viewcat.asp?cat=21&mode=gfx

2

Interesting, do you have a lot of funds available?

Unless you are using pre-determined routes/waypoints, you would need to develop something along the lines of the Urban Grand Challenge, but scaled down a bit.

darpa.mil/GRANDCHALLENGE/

3

Funding Options

Initial prototype development (funding secured)
I’m funding this with an initial budget of $2000.
This should be sufficient to assemble the first prototype.
The motors and controller seem to be the main cost. I will probably prototype multiple platforms but re-use those components.

Other funding options that I’m researching:

Grants
President Obama has announced $2.4 billion dollars in new grants for electric vehicle development.
Local government grants at the State, County, and City level might also be options.

Venture Capital
With a working prototype in hand I should be able to gain some venture capital if needed, leading up to an IPO.

Franchise
On-going funding will be provided through a franchise fee (franchise fees would probably run from $10,000 to $30,000). The franchise fee would enable use of the on-line reservation and billing system, provide tech support, and allow purchase of pre-assembled taxi robots. It would also allow franchise owners to partner with other franchises to share dispatch operations during peak times or to provide coverage during off hours.

Exhibitions

“Purchase a ticket to ride a robot!”
This would be done at County Fairs and Shopping Mall Carnivals. Think cones and other obstacles to maneuver around quickly in a parking lot, maybe some curvy line following too. All relatively safe for the kids.

Bumper Cars
I’ve also considered bumper cars. You could compete against a robot in the ring. Would be good for stress testing the designs in a controlled environment. Most controllers support autonomous and joystick operation.

4

I’ve strongly considered pre-determined routes/waypoints.

Routes to and from sporting or concert events to predefined special event parking areas.

Safe ride that provides transport at night on large campuses.

Many parks are not fully handicap accessible. A ruggedized robot rental could take you on a guided tour that a standard electric wheelchair wouldn’t be able to traverse.

I’ve considered pulling other cars (aka tram) for things like Zoo tours.

Package Delivery in large business parks for inter departmental mail. The robot would be able to drive from set way points and enter a building for delivery.

My hope is that I can use curb detection and road detection along with GPS to stay on track. Additionally dispatcher/operators would be monitoring and could help a robot out of a complex situation and get it moving again quickly. Robots approaching a busy intersection or a section of road that is known to be problematic would sort to the top of the observation priority list to assist in closer observation. I’m hoping a robot could complete an entire block on it’s own without intervention by a dispatcher. I’d like a dispatcher/operator to be able to handle at least 5 robots operating simultaneously.

I’m also going to experiment with sidewalk and bike trail tracking algorithms.

Robot use on streets would be limited to low traffic times. Transporting elderly to doctor appointments or the grocery store during none rush hour traffic times might be a good use case.

Operating in small towns with a population of 4,000 or under (where taxi services are not normally available might be another consideration). A municipality might partially subsidize this type of service.

I’m not ruling out in home use for disabled people that have limited mobility for operating a regular wheelchair. Voice operation and a touch screen and joystick would probably be added options. The dispatch/operator could still be used for remote control assistance when operating the vehicle outdoors.

5

Speaking of DARPA. Did you see this?

Discovery Channel Monster Garage.

roboteq.com/index.php?option … &Itemid=87

The steering motor in this example looks to me to be very much like a giant servo!

http://www.roboteq.com/images/modepositionloop.gif

"In the closed-loop position mode, the axle of a geared down motor is coupled to a potentiometer that is used to compare the angular position of the axle versus a desired position.

If the potentiometer senses that the motor has not reached the desired position, power will be applied to the motor until that difference becomes zero. A PID filter is incorporated inside the controller to ensure that the motor moves as quickly as possible to the end destination without overshoot and/or oscillation"

How much would you pay to ride on top of a giant Phoenix at the fair? Now if someone could help me wire this up to an ssc32… 8)

6

He controls GM so it seems a waste to me to pump even more taxpayer money in to this type of research when billions of taxpayer money are already pumped into GM, probably indefinitely. :open_mouth:

7

easy on gm sn69 i might not have a job soon if they don’t get going in the rite direction. but as far as spending more money in r/d its the government and big oil corps that have held us back. but most of the cool things you see in cars to day were developed by Oldsmobile did you know that Oldsmobile patented a hydrogen engine that ran on water back in the early 80 but the government did not allow production

8

sorry but if the gov’t did not allow production then it is because someone paid the individuals in the gov’t enough money to prevent it from happening.

9

Perhaps you can show a link to the patent to which you speak. I think you are perhaps just “easy” when it comes to urban legends. :wink:

en.wikipedia.org/wiki/Water-fuelled_car

10

ok ok i may be wrong about the patent but do you really think we will see any other source of fuel until there is no more petrol

i read about the Oldsmobile hydrogen car in the book fuel from water at barns and Nobel

and that it rite there the oil corps

sorry for hijacking the tread br549 :cry: :blush:
lokiwizz runz away with tail tucked between leggs

11

Not saying I like government buying out everyone, having said that…
From CNN
“Len Blum, managing director of Westwood Capital, an investment bank, said he thinks there’s a better chance for taxpayers to recover money from their investment in GM than with insurer AIG. Taxpayers also have a majority stake in that company, and the government has kicked in more than $180 billion so far to keep AIG afloat.”

12

Bumper Cars
After more discussion with the neighbors the bumper car concept seems to be gaining the most momentum.

Pros
Will help us to design, test, and demonstrate a safe design
Will help us find any weak points in the robot design
Will allow us to do real world robot observation and test avoidance and tracking algorithms
Do all this while making money that can help to purchase more hardware to get the taxi service off the ground
The web based reservation (ticket) and payment system could be implemented under this use
The web based sensor data monitoring system could be implemented under this use
Most work done during this phase can be directly applied to our taxi design

Use Case
Novelty, get in a ring and try to out drive a robot!
You would bump against your friends and robots in the same “ring”.
Riders could download pictures or video streams after the ride from the web, along with sensor data such as speed, number of bumps, GPS path, force from impacts, picture of you and the robot or your human competitor at impact.
Easy setup in any parking lot.
Locations could be County Fairs, Shopping Mall Carnivals, kids parties.

Setup
A parking lot with coned off area.
Virtual walls keep the robots in the coned off area
store.irobot.com/product/index.jsp?productId=2731666&cp=2804606.3335976&parentPage=family
One of these could be placed every 13-feet all the way around the coned off “battle field”.

Robot requirements
Robot bumper cars would be equipped with a joystick or could run autonomously.
Robots would have different colors. Blob recognition would be used for the robot to determine which bumper cars to pursue or avoid.
If the drive wheels are 14 inches the caster wheels should be 6 to 8 inches and spring loaded
A bumper design that would surround a round platform. Design might be a rubber like bladder or inner tube or round rubber ring with springs

13

Seats - for bumper cars
Something along the line of a go cart or racing seat.
sstmotorsports.com/Ultra_Shield_Intermediate_Racing_Seats_p/usis.htm

14

Taxi Cab
Looking for off the shelf cab solutions. Many cabs have lights, wipers, and heaters.
curtiscabs.com/CabsAndAccessories/default.aspx?ID=4&name=div1

15

Phase 1 - EV build

Phase 1 is pretty much complete (of course there will always be tweaking). I drove my “robot” at the hybrid fest last weekend (got to lead a $100,000 Tesla into the show) and I’ve been driving it to work.

I’m starting Phase II Semiautonomous Operation

I want to mockup a 1:8 scale model. I was thinking about keeping it pretty simple. Just the frame and no seat. The battery pack area could house the BAP and/or SSC-32 (and batteries). It would have three tires and a servo on the front for steering (probably no handlebars), and a continuous servo on the back to drive the rear wheels. I would also like to have a disc brake. I was thinking about just putting a washer on the rear axle and having another servo that pushed against it to stop it.

The goal would be to test the software on the 1:8 scale model for initial testing and then run the same software on the full scale trike. I’d like to sell the model on my website for anyone that wanted to contribute to or test the software.

My goal is to create an affordable open source autonomous vehicle platform. I’ve already talked with a local trike builder and they said they could build the frame.

What I like about this platform is that it’s small enough to fit in the back of my PT Cruiser for transport and could easily be assembled in a dorm room for initial build and testing on jack stands.

http://www.otherrobots.com/lynxmotion/PTCruiserTrike.jpg

I was hoping I could find someone that could mockup the 1:8 scale model and possibly produce it for me.

http://www.otherrobots.com/lynxmotion/trikefram02.jpg
http://www.otherrobots.com/lynxmotion/trikeframe.jpg

Again the frame should be simple. I want to keep the cost down but have it functional to test the software. Just three pieces of metal to form a triangle (I even thought about a triangle piece of metal) as the base. Something to mount a servo at an angle to a front tire (for steering) and one on the back for a continuous servo. Then some kind of mounting for a BAP/SSC32 and a battery pack. It doesn’t have to have shocks and handlebars or seat mounts etc.

The trike is about 53" x 72". I can post the exact dimensions if anyone is interested in giving it a shot.

16

I don’t know about the scale off hand, but take a look at this thread:

lynxmotion.net/viewtopic.php?f=9&t=5593&hilit=micromouse

KM6VV

17

Something like this might work.

http://www.np.edu.sg/alpha/nbk/umouse/images/Min5.jpg

I would have to set the tires to the back and mount one on the front for steering. I don’t want tank steering so I would only have one servo powered wheel in the back. I like the gear on the back side of the tire and how it hooks to the servo. I might try to find a wheel like that for the drive wheel.

Thanks

18

Yes, that’s a nice 'bot!

However, the motors are over $100 USD as I recall. there would be much design work (PCB) for you to do to make you’re “bicycle model” steerable “trike” out of it.

Both the trike and a differentially steered 'bot (two drive wheels, one castor) are considered to be a “bicycle model” 'bot. Basically the bots can turn in their own radius. Slightly different drive algorithm, but same results.

Alan KM6VV
(wish I could find the reference for “bicycle model” I was reading).

19

My new VBL1650 controller arrived. This is really the start of my dream/obsession with having a robot I can ride on. The VBL1650 accepts commands received from an RC radio, Analog Joystick, wireless modem, or microcomputer.

And so the journey begins…

Picture of new controller next to my blown Kelly controller for comparison.

http://www.otherrobots.com/endlesssphere/roboteqcontroller.jpg