Maya Cakmak

Instructor

[eeelaineeewxy]

Project idea: Robot for Pouring

People with impaired vision have problems with determining how much liquid is in a cup. Without the help of any technology, they often put their fingers into their glass to determine when to stop pouring the liquid into the container or ask a caregiver to complete this task for them. (Source: 14 Unreal Things You Never Realized Blind People Have To Deal With)

Since drinking water is an essential task people perform multiple times every day, having a robot that pours liquid for them is helpful, especially when the liquid is hot. With the help of a robot, people will no longer need to wet their fingers each time they try to pour some liquid, or potentially get scalded. This will also save time for caregivers. Let’s think of the scenario when it is at the middle of the night, and a visually impaired person is thirsty, if he can manipulate a robot to pour liquid for him, he will no longer need to wake up a family member or mess up the counter. By completing this daily task on their own, visually impaired people can get a sense of independence.

We can use a Fetch robot to perform this task. In order to accurately determine when to stop pouring the liquid, we also need to make special containers (such as mugs for drinking water and bowls for breakfast cereals) to accompany various tasks. The robot also needs a peripheral to get the type of liquid and the amount to pour from the user. This can either be a speech recognition device or a button panel with Braille. Due to various sizes and weights of liquid containers, determining how to lift the container and pour out the liquid from a container is a challenging task for the robot.

There are some existing solutions to this problem such as liquid indicators and the Braun Bell Mug. A liquid indicator is a kind of device which after being attached to the glass, can vibrate or buzz when the liquid approaches the glass top.

(Image of a liquid indicator. Source: Liquid Level Indicators, Liquid Level Sensors)

The image below shows the instructions of Bell Mug. According to Andrew Liszewski, “Inside the mug are three sensors that can detect how far it’s been filled, and the user simply has to press one of three buttons on the handle to choose what level they’d like. When the desired amount of liquid in the mug has been reached, an audible alert will sound letting them know they can stop pouring.”

(How to use the Bell Mug. Source: Braun Bell Concept Mug For The Visually Impaired)

In both of these solutions, the user still has to perform the task of pouring the liquid on their own, so that they still might spill the liquid out if they fail to aim at the cup or over-tilt the container. Using a robot instead can make this job more pleasant, free the user from manual work, and prevent the user from getting scald by the hot liquid.

As long as all the containers are stored in places accessible by the Fetch robot and the user is able to give instructions described above, this solution is feasible.

[oursecretstudenttreethings]

Robotics Capstone (Project Ideas)

Problem

Do you remember the time when you traveled with your friends but had to ask strangers to take photos for the squad? Or maybe you happened to be traveling alone but had to struggle to take a photo for yourself? Frankly, there are a few solutions, but none of them is perfect. You can use a selfie stick, but it’s dangerous and not that easy to use. You can use the wide angle selfie cameras on Google’s latest Pixel phones, but it would distort the picture. If only you could have someone to follow you around and take pictures of you wherever you please but a camera man is just too expensive to hire……

Solution

Here is where a picture-taking robot comes in. We propose to program the Fetch mobile manipulator robot to take pictures of designated users. Using a robot has several advantages: 1) it costs little, only for charging and some infrequent maintenance when needed; 2) it’s safe to use, and the user’s phone is unlikely to fall from robot as opposed to a selfie stick; 3) it takes great pictures! Its arms are long and stable and can hold the camera steadily at any position the user wants

The robot is also very feasible. We plan to build it so that the robot can hold a picture-taking phone in its hand, adjust the phone position with the arm and “eyes”, and finally pressing the take-picture button by increasing the force of holding the phone (suppose there’s a button on the phone responsible for picture taking). This process can be controlled by a web interface on another user’s phones, or by hearing the phrase “Take a picture”. None of the above steps requires special or unavailable technology, and we’ll have all the technology we need to build the robot from day 1.

There are two business models for the robot: 1) tourist sights can purchase and deploy some robots at popular places for travelers to use; 2) users can purchase or rent a robot to follow them and take pictures of them wherever they go. Both of these simple models are easily profitable and simple to execute. Given the fact that this problem is widely present and currently has no better solutions, our robots are going to bring us a great deal of fortune while keeping the users smile with their best photos.

[rrroobit-blog]

Startup Idea: Expired Grocery Finder Bot

Problem: Grocery stores have a need to remove expired items from shelves so that customers do not end up buying them. In order to do so, they have employees manually check items when restocking. However, most stores are far too large relative to how few employees they have to efficiently check all possible items on the shelves. Under staffing, combined with the near impossibility of keeping track of what items have and haven’t been checked, often leads to expired items being left on the shelves. Customers complain upon finding those items, and employees are punished as a result.

Solution: I propose that the Fetch bot be programmed to detect and removed expired items from shelves. Each item will be clearly labeled with their item number and expiration date, using either a visible barcode or an RFID tag. Since items with more recent expiration dates are always placed in the front of the shelves, the robot will only need to scan the items in the front of the shelves in order to find  expired ones. The robot will move through the isles, scanning items as it goes. Upon finding an expired item, the robot will look up the item number in its database to determine how to pick it up. Afterward, it will place it in its collection bin. If the item is too large or heavy to be picked up, the robot will instead notify an employee of the presence of the item.

Evidence for the problem and the feasibility of the solution:

Myself being a frequent customer of grocery stores, I had previously run into expired items on the shelves. I had even accidentally bought some, due to having failed to check the expiration date. To help find support for the importance of the problem, and see if my prior experiences were accurate representations of the current existence of the problem, I decided to go to my local Safeway and see if I could find any expired items.

Within a minute of entering the store, I found one:

This picture was taken on 4/1/2017, one day after this item’s sell by date. This clearly shows that the current process for removing expired groceries is not as efficient as it should be. Another thing it shows is that pastries in particular are labeled manually by the store itself, and are labeled very clearly on the packaging. This means if a store were to adopt the Fetch robot to manage expired pastries, it would be very easy for them to produce labels for their items which the robot could use.

—

To find further evidence to support the existence of the problem and feasibility of my solution, I decided to interview a friend of mine who used to work at a Safeway. I asked him what the process was for handling expired items. What he told me was that whenever they restocked the shelves, they were responsible for checking that all items currently on the shelves were in date. They would rotate old items with more recent expiration dates to the front of the shelves (evidence for feasibility: robot can scan items on front of shelves without having to dig). According to him, the expired items most likely to end up being missed were the ones which people didn’t buy and thus did not need to be restocked and reshuffled by hand. They would do a cursory check of the shelves, but when new items didn’t need to be rotated in, expired items would sometimes be missed (evidence for the problem: expired items remain on the shelves). He described the produce department where he worked as “understaffed” (evidence for the problem: robots can help alleviate workload, work 24/7 while people can’t), and according to him, they would receive complaints from customers about expired food items at least once a week (evidence for the problem: expired items are missed and inconvenience the customer). His boss would “chew out the whole department” whenever it happened, despite the fact that their department had “way too few people” to handle restocking efficiently (evidence for the problem: the job is made worse because the staff can’t eliminate the problem).

[rrroobit-blog]

Startup Idea: Dish Pickup Bot

Problem: Restaurants, both with and without waiters, have a need to pick up dirty dishes from tables. Restaurants without waiters rely on customers to dispose of their dirty dishes themselves. This creates two problems: customers which don’t pick up after themselves leaving behind their dishes, and other customers being inconvenienced by the need to take care of their dishes themselves. The first problem creates excess workload for employees, and possibly affects the experience of customers which just entered the restaurant and are looking for a place to sit. The second negatively effects quite possibly every customer who enters the restaurant, albeit in (usually) a fairly minor way. In restaurants that do have waiters, a different problem arises. When one goes to a restaurant with friends, and is engaged in conversation with them, a waiter walking up to pick up used dishes can be intrusive. A waiter walking up and asking for your dishes while you are in the middle of a personal conversation will generally interrupt the conversation, and require you to communicate with the waiter rather than the person you’re speaking to.

Solution: I propose the Fetch robot be programmed to travel to tables around a restaurant and pick up used dishes from them. The process can be triggered by an employee of the restaurant, who upon noticing a table has dishes which need to be picked up, inputs the table number and tells the robot to go there. If the customer is still there, they can hand the dishes they no longer need over to the robot, who will take them back to be cleaned. If not, the robot will simply pick up every dish currently on the table. A robot would both help alleviate the workload of the restaurant staff, and would be far less intrusive than a person while picking up dishes.

Evidence for the problem and the feasibility of the solution:

My first context comes from that of myself as a restaurant patron. When I go out to eat with friends, we generally want to sit together and talk while we eat. However, on more than one occasion, waiters have ended up interrupting our conversations so that they might pick up our empty dishes. My friends and I were forced to stop speaking to each other and interact with the waiter instead. A robot would be much less intrusive, since there would be no awkwardness associated with continuing a conversation in the presence of a robot like there would be if you kept talking in front of a stranger.

I’ve also experienced a different problem when dining alone at restaurants without waiters. Many such restaurants require the customer to pick up after themselves and drop their used dishes off at a designated area. In smaller restaurants, this generally has not been a problem, since there are only a few exits, and the dish drop-off is placed near them. However, at certain bigger restaurants, such as Local Point at Lander Hall, this is not the case. Here is a picture of Local Point’s dish drop off area relative to the rest of the restaurant:

Keep in mind that not all of the restaurant is visible in this picture. There is a quite a bit more well out of frame. As you can see from the picture, the dish drop off is quite a ways away from the exit, as well as many of the seats. What this means for customers is that when they are done eating, they must walk all the way from their seat to the dish drop off, then all the way back to the entrance. Due to the size of the restaurant and inconvenient location of the drop off, this is slow and a poor experience for the customer. A robot picking up the dishes would greatly reduce the distance the customer has to walk when they are ready to leave the restaurant.

—

After reflecting on my own experience as a restaurant patron, I began to wonder if other customers shared my same views. I decided to go to yelp.com and look for restaurant reviews. In particular, I searched for reviews that mentioned the words “service”, “waiter”, and “interrupt”. I also manually browsed the 1 and 5 star reviews to see if any patterns emerged. The following are some excerpts from reviews I found for the restaurant Canlis in Seattle:

“But the service was what made our evening. Our wait staff were there when needed without ever interrupting our dinner.” (evidence for feasibility: lack of interruptions is valued highly by customers, and reducing interruptions would thus increase customer satisfaction)

“Our server spoke loudly. So much so, any interaction at our table or tables around us interrupted our flow of conversation” (evidence for the problem: interruptions reduce customer satisfaction)

“It felt like we were served by thirty different people – four different people filled our water glasses alone! A ridiculous amount of staff, especially since this was a Valentine’s Day dinner and I had hoped for something more romantic and intimate. We were constantly being interrupted by someone new. I swear, there are probably well-dressed homeless people who just wander in off the street and unobtrusively join the waitstaff – that’s what it felt like. It’s like we were assailed by an army of servers. By the end of the evening I felt like I was clinging to our main server, because she was a familiar face and one of our only servers who remained constant throughout the evening. Hoards of waitstaff were very distracting. It felt like being attacked by shoe salesmen during Nordstrom’s semi-annual sales.” (evidence for the problem: being interrupted, especially by unfamiliar people, was distracting and took away from the “intimate” atmosphere)

Overall, I found that quite a few people shared my views on being interrupted by waiters. Some of it had to do with their behavior, such as speaking loudly, while some of it had to do with being forced to interact with unfamiliar people. A robot would not cause either of these problems, which could definitely help improve the customer’s experience. I also read reviews for several other restaurants, and noticed that low ranking reviews for restaurants tended to mention slow service:

“The service between the apps and salads were really slow… and again slow between the salads and the entrees. They even had a "bread” course where they went around serving bread too! Which added to the slowness of the service - it would have been much more efficient to have the bread on the table - after they served the bread, they then came around with the butter, which by that time we had already finished the bread because we were soo hungry!“ (evidence for problem: slow service leads to dissatisfaction)

"Service was also slow with an average time of 15-20 minutes between each course. We ate too much bread and butter due to the delays.” (evidence for problem: slow service leads to dissatisfaction)

Part of the workload being done by a robot could help improve the speed of the service. While it would not solve bottlenecks at the kitchen, it could help solve a shortage of waiters.

 - Image of Canlis, Taken From Yelp

[koar8470-blog]

Start up idea: Clothes store pickerbot

Problem: There are many fitting rooms at large fast-fashion stores such as H&M, UNIQLO, and Gap, and usually 1 to 3 staffs are there to return items back to where they were and bring items of different sizes for customers. For staffs, it’s time-consuming to go back and forth between the fitting room and clothes lacks while responding customer’s various demands. For customers, it’s uncomfortable to ask staffs to change the size of item again and again.

Solution: I propose to program the Fetch robot to return/get items of different sizes to/from clothes lack to help customers to find the size that fits them best easily. The clothes lack will be designed so that the clothes can be readily reached by the robot. Hanger lacks would be good. The order by the customers are taken by a tablet attached in the fitting room, and it will be sent to the robot. Easy tasks, like just trying one size up, can be done by voice command.

Evidence for the problem and feasibility of the solution: I interviewed several customers and a staff at UNIQLO at the Bellevue square. Here are some selected quotes from them:

“It’s troublesome that I have to wear my own clothes back on every time I try other size or item because there may be an interaction with a staff. It’d be nice if a robot can bring other clothes for me, and I can stay naked in the booth.” (Customer A)

“I enjoy talking with staffs and having comments from them, so I don’t need a robot.” (Customer B)

“There are so many tasks and human resource is never enough. The tasks include getting different sizes/items, folding tried clothes and return to the lack, taking orders of alteration, checking the stock, and so on. If a robot could help any part of these tasks, I could use more time for customers.” (Staff)

There are certain demands for the pickerbot, although some of the customers feel don’t need it. However, if robots are used in the workplace, staffs would have more time per customer. As a result, the robot is beneficial even for those who don’t find the need for it for now.

The average wage for the staff is approximately 14 dollars. So if 2 staffs are working at the fitting room constantly, it costs 2*14[dollar]*8[dollar/hr] = 224[dollar/day]. Assuming the store is open for 360 days a year, the yearly running cost is, 224[dollar/day]*360[day] = $80,640. Therefore, the robot can be installed for the cost of two people for one year. Considering that the robot doesn’t need as much cost after the initial investment, this solution is very feasible in terms of the cost efficiency.

The store design may need to be optimized for robots. IKEA is a good example. There may be sample exhibition area and stock area. The samples are hanged on racks and the location of the items and racks are programmed in robots so that they are easily reached by robots. Once a customer decides to purchase an item, the robot goes to the stock area and move the item to a bag. When the customer finishes shopping, he/she can pick up the bag and pay at the register. The list of items is already stored in the database, so it’s quick.

[mayacakmak]

7Bot has shipped

We have received our 7Bot last week, although we are still waiting on the software packages for it.

In the mean time, Dan Butler, a graduate student at UW CSE, has created a simple software that allows saving a sequence of poses and suction cup states that can be replayed by the robot. His GitHub repository is here. Let Maya and Sarah know if you would like to use it for your projects--apparently it could easily lift a soda can or a candy bar ;)

[mayacakmak]

Aido robot is great @cse481sp16news

[mayacakmak]

Test

Hello world..